Arrowhead Tunnels Project Surface Water Impact And

ARROWHEAD TUNNELS PROJECT SURFACE

WATER IMPACT AND RECOVERY

ASSESSMENT

Specialist Report

October 2012

Neil Berg

TABLE OF CONTENTS

EXECUTIVE SUMMARY 3

INTRODUCTION AND OBJECTIVES 5

FACTORS AFFECTING STREAM AND SPRING FLOWS IN THE ATP AREA 6

SURFACE WATER MONITORING HISTORY AND APPROACH 6

SURFACE WATER REGIMES 10

SURFACE WATER MITIGATION HISTORY 14

METHODOLOGY 14

APPROACHES TO IMPACT AND RECOVERY ASSESSMENT 16

SITES ASSESSED FOR POTENTIAL CONSTRUCTION IMPACT 20

RESULTS 21

IMPACTS TO TRIBAL LANDS 49

ENVIRONMENTAL EFFECTS TO SURFACE WATERS FROM ONGOING OPERATIONS AND MAINTENANCE OF THE ARROWHEAD TUNNELS PROJECT 49

LESSONS LEARNED 50

CONCLUSIONS 51

EXECUTIVE SUMMARY

The Arrowhead tunnels are part of a larger, 48-mile water transmission facility (the Inland Feeder Project) incorporating pipelines and tunnels between Devil Canyon in the northwestern San Bernardino Mountains, and Diamond Valley Lake near Hemet, California. The objectives of the Inland Feeder Project are to provide a six-month water supply within the greater Los Angeles basin in case a catastrophic earthquake impairs the ability to move northern California water across the San Andreas fault zone into the Los Angeles basin. An additional objective of the Inland Feeder Project is to improve water quality compared to water from the Colorado River aqueduct.

The two Arrowhead tunnels total approximately 9 miles in length stretching from City Creek on the south to Devil Canyon on the west on the lower slope of the San Bernardino Mountain Front Country. Tunnel construction can inadvertently ―leak‖ water from groundwater aquifers into the tunnel as the tunnel boring machines advance through the mountain. This groundwater would potentially have linked to the ground surface to provide source water to springs and streams. By removing groundwater tunnel construction can reduce surface flow or completely de-water streams and springs. This potential de-watering is a crux of Forest Service concern about the Arrowhead tunnel project.

Objectives of this assessment are to identify (1) surface water assess impacts (if any) from construction of the tunnels and (2) identify the level of recovery in surface waters from any impacts.

Approximately seventy-five locations in the Arrowhead project area were monitored for surface water flow, typically beginning in 1994, at intervals from weekly to twice annually. Measurements at almost all locations were made instantaneously, using either a current meter or by direct volumetric determination. At two locations in lower Sand Canyon automated gaging provided flow measurements at 10 or 15- minute intervals. These more detailed measurements proved critical in quantifying effects of summer evapotranspiration that caused flow changes up to 50 gallons per minute over 24-hour periods.

Two related procedures were used to assess potential tunnel construction related impacts to surface waters. Both compare monitored flows at a reference (control) site with monitored flows at sites potentially impacted. The first procedure is an ocular comparison of flows through the entirety of the data record. A visual change in the flow relation between the two sites at or about the time of documented construction-induced impact at a proximate well(s) indicates a good likelihood for a construction-induced impact at the surface site. The second, more detailed, approach compares flows only for summer ―baseflow‖ periods when groundwater is the only source for surface water. A relationship is generated for the pre-impact/baseline period (determined by the date of impact at one or more nearby well(s)). A change from the baseline relationship indicates a high likelihood of a construction-induced impact to surface waters.

Flows at eighteen monitoring locations (including sixteen on NFS lands) were reduced at one time or another as a consequence of tunnel construction. In addition, flows on the San Manuel reservation at the base of Sand Canyon were also reduced. Mitigation, as irrigation at selected locations above many of the impacted sites, was routinely applied for several years. The irrigation effectively compensated for much of the construction-caused reductions in natural flows. As of summer 2012, flows at sites 48, 53, and 636, all in upper Sand Canyon, sites 56, 58, and 181 in a tributary to City Creek, and sites 45 and 154 in Borea Canyon continued to be barely within or below the anticipated natural flow ranges.

There are no anticipated environmental effects to surface waters in the project area anticipated with routine operation and maintenance of the Arrowhead Tunnels (i.e., the 2012 Special Use Permit). The

tunnels are steel-lined and consequently no groundwater should leak into the tunnels or otherwise be diverted from natural linkages to surface water sites. Because of potential residual surface water effects from tunnel construction it is anticipated that at most monitoring of recovery of stream and spring flows would be needed through 2013 at locations in Sand Canyon and City Creek tributary sites 56/181/58. Because no direct or indirect effects are anticipated there will be no cumulative effects stemming from tunnel operation and maintenance.

INTRODUCTION AND OBJECTIVES

The Arrowhead Tunnel Project (ATP) comprises two 12 ft-diameter water supply tunnels totaling 9 miles in length under National Forest System lands in the San Bernardino National Forest. Tunnel construction started in 1998 and was completed in late 2009. Major potential environmental effects of the ATP include depletion of groundwater aquifers—as tunnel mining intercepts and removes groundwater from aquifers—and de-watering of springs and streams that have groundwater as their water source. Indirect effects include potential degradation of riparian and aquatic habitat for a variety of biological species, including federally-designated Threatened and Endangered and State or Forest-sensitive species.

Per a variety of USDA Forest Service (FS) and other federal and state agency guidelines and directives, the San Bernardino National Forest needs to assess potential impacts and work to restore affected ecosystem components. Specifically, for instance, FS Interim Directive 2020 (Ecological Restoration and Resilience, 9/16/08) defines restoration as: ―… the process of assisting the recovery of resilience and adaptive capacity of ecosystems that have been degraded, damaged, or destroyed. Restoration focuses on establishing the composition, structure, pattern, and ecological processes necessary to make terrestrial and aquatic ecosystems sustainable, resilient, and healthy under current and future conditions.‖ This definition drives FS staff to ―… re-establish and retain resilience … to achieve sustainable management and provide a broad range of ecological services.‖ In this context identification of potential hydrological and biological effects is a first step in both the mitigation of current effects and the restoration of any effects continuing after the completion of a project.

The focus of this report is surface resources, broadly interpreted to include riparian and aquatic-dependent biota and surface water flows themselves. The premise is that surface flows are a primary driver of the health of the aquatic and riparian biota and that potential impacts to biota are confined to areas where there are surface water impacts. More specifically, this report addresses spring and stream flows. Potential biological effects are addressed in separate documents. Although effects could be and probably are somewhat broadly spread in at least some project area catchments, the historical database for surface flow magnitudes is limited to between two and eleven surface water monitoring locations within each of six primary catchments (Figures 1a and 1b). Hence this impact assessment is effectively limited to a maximum of approximately seventy individual surface water monitoring locations.

Three primary objectives are to (1) identify surface waters impacted by tunnel construction activities (status/presence-absence), (2) assess the confidence of the impact identification (e.g., firm or ―grey area‖), and (3) evaluate current impact trend (e.g., flows rebounding/recovered after impact or no evidence of rebound). This report does not address potential water quality effects because they have been assessed elsewhere (e.g., Berg 2008), including evaluation of potential effects of chemical leaching from mitigation water conveyance pipes, elevated chlorine concentrations in the mitigation water from the use of domestically-treated water, and elevated mitigation water temperatures due to water heating in the conveyance pipes. No water quality effects of significance were identified. This assessment incorporates information available through mid-2012.

Another premise of this report is that tunnel construction can affect groundwater dynamics but not precipitation regimes. Therefore intermittent and particularly ephemeral surface water regimes that do not include summer flows are less important because they are relatively less influenced by groundwater. This distinction means that the focus of this assessment is on perennial streams which in the Mediterranean climate of the project area flow in summer due only to groundwater sources.

A primary aim of this assessment is to identify surface waters impacted by tunnel construction. ―Candidate impacted site‖ is the term that is used in this assessment to identify surface water monitoring sites that are potentially impacted by tunnel construction.

FACTORS AFFECTING STREAM AND SPRING FLOWS IN THE ATP AREA

The climate of the ATP area is typically classified as ―Mediterranean‖, with precipitation occurring primarily from late autumn through early spring. A major ramification of this precipitation regime is the lack of precipitation inputs to surface flow in the summer months. During summer low-flow periods groundwater is the primary source for surface flows. In that tunnel construction can ―capture‖ groundwater and remove it from natural aquifer systems, the coupling of the Mediterranean climatic regime with potential groundwater removal during tunnel construction is a major potential hydrological impact of the ATP.

A variety of meteorological conditions occurred throughout both the baseline and construction periods including severe dry conditions from 2000 to2004 (less than 15‖ annually each year) and high precipitation during the winters of 1997-1998, 2004-2005, and 2010-2011 (up to 40+‖ annually). In addition, locations near the project area can experience some of the highest intensity precipitation in California, with 10-inch totals during a 24-hr period recorded several times historically at Lytle Creek Ranger Station, Lake Arrowhead, Big Bear Lake Dam, and Crestline. Several of these locations are in the highlands of the San Bernardino Mountains, a potential source area for regional recharge of groundwater aquifers in the ATP area.

Besides precipitation variability, other factors influenced surface water flows during the project. Specifically, a major wildfire during autumn 2003 denuded the landscape over the entire project area. Massive flooding and erosion occurred the following winter when over 8‖ of precipitation fell during a 24-hr period on the Waterman Canyon portion of the project area (Psomas 2004). This fire-flood cycling is common in southern California (Cannon and DeGraff 2009, Ainsworth and Doss 1995) and was recognized there as early as the 1930s. It results when the burned off vegetation can no longer hold soil after intense rainfall occurs. The subsequent rapid runoff and increased sediment are washed down stream channels. One ramification of this phenomenon is the highly dynamic nature of channel morphology; depending on the stage of the erosion-sedimentation cycling, channels can be deluged with sediment or reamed clean (Figure 2a and b). This dynamic nature of channels in the project area complicates surface water monitoring; monitoring at bare bedrock locations is strongly preferred to reduce the potential for un-measureable underflow within channel sediment. However, when channels are deluged with erosion-induced sedimentation accurate measurement of flows can become problematic.

SURFACE WATER MONITORING HISTORY AND APPROACH

At many monitoring sites typically one measurement in 1991 or 1993 was followed by more frequent monitoring initiated in 1994 or 1995. Some sites were added later (e.g., 1997 or 1998 at several locations in Sand Canyon). The monitoring focused on the western tributaries to City Creek, main stem and tributaries in Sand, Little Sand, Borea, Sycamore, Badger, and Ben Canyons with additional monitoring primarily on the main stem channels of City, Strawberry, East Twin, Waterman, and Devil Canyon Creeks. Tunnel mining began on the City Creek portion of the eastern tunnel alignment in May 1998 and in 2003 for the western alignment and the bulk of the eastern alignment. Thus many monitoring sites

benefit from at least a 16-year period of measurement including several years prior to the ―City Creek‖ construction and over 9 years prior to the start of construction on the non-City Creek tunnel segments.

Seventy-nine original surface water monitoring sites were reduced over time for a variety of reasons, including site burial by landslides and removal from the monitoring network after private landowners revoked permission to access sites. Seventy-two surface water monitoring sites were operational as of spring 2003 with the majority of these located on FS land (MWD 2003).

Sites actively monitored in 2003* are tabulated below, along with selected site-specific information through 2003 (source: MWD 2003).

Site Number Group ID1 Resource Location Range in Flow (gallons/minute) Type2 Low High Arrowhead West - Springs and Streams 8 A S Devil Canyon 0 4 153 A S Devil Canyon 0 4 190 A G Devil Canyon 0 20000 193 C G Devil Canyon 70 2100 620 B G Devil Canyon 200 10000 9 A S Ben Canyon 0.25 50 10 A G Ben Canyon 0.4 40 11 A S Ben Canyon 0.06 30 157 A S Ben Canyon 0 12 21 B S Badger Canyon 0.3 45 26 A S Badger Canyon 0 70 27 A G Badger Canyon 0 90 28 A G Badger Canyon 0 8 152 C G Badger Canyon 55 650 213 A S Badger Canyon 0.15 15 214 A S Badger Canyon 0 85 20 A G Sycamore Canyon 10 150 30 A G Sycamore Canyon 12 300 95 B G Sycamore Canyon 1 400 156 A G Sycamore Canyon 0.8 150 182 B G Sycamore Canyon 20 280 205 A G Sycamore Canyon 10 220 Waterman Canyon - Springs and Streams 17 B G Highway 18 0 120 65 B S Highway 18 2.5 40 627 B G Highway 18 0 40 93 B S Waterman Canyon 0 26 134 B G Waterman Canyon 200 47000 191 A G Waterman Canyon 30 30000 644 B G East Twin Creek 90 11200 38 B G East Twin Creek 4 23000 189 B G East Twin Creek 160 26000 201 B S East Twin Creek 0 25

Site Number Group ID1 Resource Location Range in Flow (gallons/minute) Type2 Low High 624 B G East Twin Creek 220 17000 628 B G East Twin Creek 0 3000 645 B G Strawberry Creek 40 1000 676 B G Strawberry Creek 0 650 677 B G Strawberry Creek 0 900 678 B G Strawberry Creek 0 3 679 B G Strawberry Creek 6 1100 629 B G Strawberry Creek 5 4400 Arrowhead East - Springs and Streams 154 B G Borea Canyon 25 210 155 B G Little Sand Canyon 38 2600 509 B G Little Sand Canyon 35 1100 510 A S Little Sand Canyon 0 25 637 A G Little Sand Canyon 0 45 48 A S Sand Canyon 1.3 11 51 A S Sand Canyon 0 1.8 53 A S Sand Canyon 0.02 50 54 A S Sand Canyon 0 180 103 B G Sand Canyon 10 3000 117 A G Sand Canyon 1.7 4500 185 A S Sand Canyon 0.03 20 635 A G Sand Canyon 0 270 636 A G Sand Canyon 1.2 300 622 B G South McKinley 0 170 633 A S South McKinley 0 2.5 151 C G City Creek 0 35000 520 B G City Creek 22 55000 56 A S City Creek (56 loop) 0 50 58 B S City Creek (56 loop) 0 4 181 A G City Creek (56 loop) 0 70 59 C S City Creek (60 loop) 0 80 60 B S City Creek (60 loop) 0 13 515 C G City Creek (60 loop) 0 21 630 B G City Creek (60 loop) 0 15 631 B G City Creek (60 loop) 0 60 632 B G City Creek (60 loop) 0 30 55 B S City Creek (Coldfoot) 0 100 209 B S City Creek (Coldfoot) 0 65 210 B G City Creek (Coldfoot) 0 130

1 Site location with respect to tunnel alignment: A: < 1500 ft from alignment; B: 1500-3000 ft from alignment; C: >3000 ft from alignment

2 S: spring; G: stream

* In addition two sites, 44 in Little Sand Canyon and 45 in Borea Canyon, are old adits from which surface water flows. ______

Monitoring frequency varied throughout the duration of the project, with monthly to twice-yearly monitoring early on, weekly at many locations during tunnel construction, and more recently quarterly monitoring at locations that didn’t experience a tunnel impact. As of mid-June 2012 bi-weekly monitoring by MWD staff was underway at locations in Sand, Little Sand, and Borea Canyons, with monthly monitoring at three sites in a tributary to City Creek. The weekly monitoring was more frequent than the frequency listed in the 1993 FEIR, which describes monthly measurements as the most frequent measurement timeframe.

During later years of the project FS staff measured flows at selected locations, most commonly at City Creek tributary sites 56, 58 & 181, and Borea Canyon site 45. This effort provided independent assessment of flow magnitudes and generally confirmed the measurements made by MWD staff.

Two techniques are used by MWD staff to measure surface water flows. Both techniques provide instantaneous, ―slice-in-time‖ measurements. Because these manual measurements are instantaneous they are limited in that theoretically flows a few seconds, minutes, hours, or days earlier or later could differ significantly from the measured flows. For lower flows, typically in the 0.01-20 gpm range, direct volumetric measurement of the time taken for flow to fill a container of known volume is used (Rantz et al. 1982). In this method an earthen dam is often constructed to provide a relatively still water reservoir. A PVC pipe is inserted into the dam with flow from the pipe falling into the container. For higher flows a current meter is used to determine the average velocity in the ―velocity-area‖ measurement method (Gordon et al. 1992). Wherever possible, measurements are made on bedrock surfaces to minimize potential underflow through channel sediment.

During part of the monitoring period an MWD weir and a tribal flume were in operation in lower Sand Canyon (near MWD site 117 and approximately one-quarter mile downstream just inside the reservation boundary). These devices allowed monitoring at approximately 10 to 15 minute intervals. Measurements from these devices clearly showed significant within-day fluctuations in flow (e.g., often 20-30 gpm differences between summer daily maximum and minimum flows) implying that the slice-in- time/instantaneous manual measurements did not necessarily represent a daily mean or median flow. To partially compensate for this limitation MWD staff attempted to measure flow manually at approximately the same time of day so that time-of-day would not be a significant cause of variation in reported flow measurements (e.g., approximately 90% of the measurements at lower Sand Canyon site 117 were made between 0900 and 1130, a time span that could nevertheless easily experience a 10 gpm flow change). This attempt was reasonably successful although this source of measurement variability at individual sites is definitely not zero. It’s also inadvisable to extrapolate and compare flows from the ―slice-in-time‖ measurements among different sites simply because of flow variability due to measurement time-of-day.

An associated aspect of the instantaneous manual measurement of flow is that daily minimum flows that could be biologically significant--for instance biota could suffer if flows drop to 0--are not generally quantified with the manual measurements. Depending on the time of day that the flow measurement is made, the daily minimum is probably not monitored, and at site 117 the manual measurement is actually much closer to the daily maximum than the daily minimum (e.g., from July through October 2008 the mean manual measurement was 51.8 gpm, the daily maximum—from the weir measurements—was 54.9 gpm—and the daily minimum, also from the weir measurements, was 11.5 gpm).

These technical measurement issues cannot be readily resolved. In total they suggest that a conservative approach should be taken in the use of the flow data for impact analysis.

SURFACE WATER REGIMES

The surface water monitoring locations can be classified as spring or streams. Most monitoring locations are on streams where winter flows can respond quickly to precipitation inputs, via runoff contributions or contributions through shallow groundwater pathways. Two true springs are monitored where water comes directly out of the ground (e.g., sites 44 and 45). Several other monitoring sites combine spring & stream characteristics in that they are springs having an upstream channel adding surface flow in winter to the year-round spring flow (e.g., sites 48, 185, 56, 154). The springs can be the most direct indicators of tunnel construction impact because the source of spring flow is entirely sub-surface waters that could be decreased or eliminated by tunnel construction.

In addition, surface flows in the ATP area can be classified into four regimes: perennial (flowing year- round), near-perennial (flowing year-round except for a few exceptions), intermittent (0 flow in many summers) and ephemeral (winter flow only directly resulting from precipitation, i.e. sporadically only in winter). The monitoring sites are classified below by regime. Intermittent and ephemerally-flowing monitoring sites are relatively poor indicators of potential tunnel construction impacts because their links to groundwater are minimal.

"Perenniality" status of surface water monitoring sites -- Inland Feeder Project

Sites are ordered from upstream to downstream in each drainage so that site 193 is the highest stream monitored in Devil Canyon.

Color Code

Perennial: always flow Almost perennial: less than five "0" flow measurements Intermittent: 0 flow often in summer Ephemeral: Often 0 flow, sometimes for many consecutive months

The definitions above are imperfect. "0" measured flow may miss near-surface sub -surface flow that is biologically effective. This may be particularly relevant to braided channels with abundant sediment in Strawberry and other creeks.

Stream vs. spring classification is from MWD documentation.

Note: "Q" stands for flow quantity and gpm = gallons/minute. "min Q" is the lowest recorded flow quantity.

Devil Streams Status Springs Status 193 Never 0 flow, but on 6-month measurement interval 110 0 flow during drought summers

190 0 late summer, 1 or 2 months almost every year 153 Often dry since summer 1999 620 Never 0, min Q = 164 gpm; monthly measurement 8 0 almost always since 8/99 & in prior summers

Ben Streams Springs 10 0 three times Aug to Dec 2003, no other 0 flows 157 0 5/01 to 10/05, no other 0 flows 11 0 8/03 to 12/03, no other 0 flows 9 0 7/02 to 10/02 and 9/03 to 12/03, no other 0 flows

Badger Streams Springs 0 between 1 & 7 months each drought summer (2000- 27 0 each drought year 2001 to 2003 progressively 1 to 3 months 214 2003) 28 Almost always 0, monthly measurements 213 0 during two summer drought months in 2003 11

152 Never 0, on 6-month measurement interval; min = 6 gpm 26 0 July 2001 to Jan 2003, then no more measurements 21 Never 0, last measured 7/01. Min = 0.3 gpm on 7/01

Sycamore Streams Springs 156 Never 0, min Q = 0.5 gpm, monthly 205 Never 0, min Q = 4 gpm, monthly 30 Never 0, min Q = 2.3 gpm, monthly 20 Never 0, min Q = 8 gpm, monthly 182 Never 0, min Q = 1 gpm, monthly 95 Never 0, min Q = 0.4 gpm, monthly

Sycamore/Waterman Streams Springs 65 Never 0, min Q = 0.3 gpm, monthly 17 0 during 1-4 more drought months 2002-2004 627 0 most summer months since 7/99

Waterman Streams Springs 191 0 Aug-Nov 2002 93 Many 0s 134 Never 0, min Q = 178 gpm, monthly

Strawberry Streams Springs 645 Never 0, min Q = 1.8 gpm, monthly 120 Many 0s thru 11/03, then flow each month 644 0 in Aug & Sep 2002 only, monthly 201 0 2/04 on, & in prior summers 678 0 3 months in 2000 & all 12 months 2004 676 0 between 1 & 5 months each summer 2000-2002 677 0 summers 2000-2004 679 0 Sep-Nov 2002 629 Never 0, min Q = 3.3 gpm, monthly 628 0 once in each of 3 yrs (in Aug), monthly 38 Never 0, min Q = 4.2 gpm, monthly 189 Never 0, min Q = 104 gpm, monthly 624 Never 0, min Q = 115 gpm, monthly

Borea Streams Springs 154 Never 0, min Q = 3 gpm, monthly 45 Never 0, min Q = 2.1 gpm prior to 5/05, monthly

Little Sand Streams Springs 637 Many 0s 510 Often 0 since 6/01 509 Never 0, min Q = 19 gpm, monthly 44 Never 0, min Q = 2.2 gpm, monthly 155 Never 0, min Q = 5 gpm, monthly

Sand Streams Springs 635 0s mid-summer months 2002, 2003 & 2004 48 Never 0, min Q = 1.2 gpm, monthly 636 Never 0, min Q = 0.3 gpm, monthly 54 0s summer 96 & 04 only, monthly 117 0s mid-summer months 2002 & 2003 53 Never 0, min Q = 0.02 gpm, monthly 103 0s mid-summer months 2002 & 2003 51 Mostly 0s since 6/01 185 0 summers 2002 & 2003

City Crk Streams Springs 210 lotta 0s since 7/99 55 0s most summer months 1999-2004 630 Mostly 0 after 6/99 209 Many 0s after 7/99 515 0s summer 02 & 03 & all 12 months of 04 59 0 most summer months 1999-2004 632 0s summers 99 & 00, & many 0s after 7/01 60 All 0s since 2/99 151 4 summer months w/ 0s entire monthly record 58 5 0s during summer months 2000, 2003 & 2005 631 0s summer months 99-04 56 No 0s until dewatering 6/00 181 0 on 2 summer months before de-watering 633 10 0s total 8/99 to 01. Record ends 2003 622 0s much of 01, 02 & 03 626 Lotta 0s since summer 1999. Record ends 2003 625 Never 0, minimum = 11 gpm, monthly 520 Never 0, minimum = 5 gpm, monthly

SURFACE WATER MITIGATION HISTORY

To help sustain aquatic and riparian ecosystems within the project area mitigation, as ―irrigation‖ water, was applied at various locations and at various times as tabulated below:

Site Duration Gallons 56/181 -- City Creek 1999-2003 4,676,076 48 – upper Sand tributary 2007-2011 11,537,653 Terrace 1 – upper Sand, west 2007-2008 653,213 Terrace 2 – upper Sand, west 2007-2008 529,109 Terrace 3 – upper Sand, west 2007-2008 567,807 Terrace 4 – upper Sand, east 2007-2008 756,231 Terrace 5 – upper Sand, east 2007-2008 556,993 Terrace 6 – upper Sand, east 2007-2008 695,784 636 – upper Sand tributary 2007-2011 7,078,112 54 – middle Sand tributary 2007-2010 3,584,995 53 – middle Sand tributary 2007-2011 2,835,396 117 – lower Sand main stem 2007-2010 14,776,717 510 – upper Little Sand tributary 2006-2009 2,225,344 45 -- upper Borea spring/adit 2005-2010 4,740,923 Borea main stem between 45 & 154 2008 124,000 17 – private, between Sycamore & Waterman 2005-2007 141,800 65 – private, between Sycamore & Waterman 2006-2008 421,340 156 – upper Sycamore main stem 2006-2008 1,182,660

Total 57,084,153

To put the irrigation amount into a broader perspective, it is approximately 4% of the total amount of water that moved (leaked) from the groundwater aquifers in the project area into the Arrowhead tunnels (not including the 1998 to 2001 City Creek tunnel inflows) during the construction phase of the project.

To supply irrigation water an infrastructure was established including the placement of six water storage tanks, several miles of plastic pipe to move irrigation water from the tanks to the irrigation points, and facilities for pumping water over 500 vertical feet up the slopes of the San Bernardino Front Country.

METHODOLOGY

Several analytical techniques are available for identifying construction-impacted surface flows. These include the use of flow duration curves (Searcy 1959, US EPA 2007), analog year comparisons, and other procedures. However these techniques generally incur relatively low precision in their predictions of impact magnitudes, are restricted to relative (versus absolute) determinations of impacts (e.g., flows are lower at a candidate impacted site rather than flows are XX gallons per minute (gpm) lower), or require relatively complicated computations.

For a variety of reasons the approach used herein compares flows at reference (or control) sites against flows at candidate impacted sites in a ―before and after‖ setting. The premise is that actual reductions in flows at candidate impacted sites can be documented in comparison to a pre-impact baseline relationship 14

between flows at reference and candidate impacted sites. The reference-impacted approach has been widely used in hydrologic assessments (e.g., USGS 2004, Lisle et al 2007). In addition to the reference- impacted flow comparison, assessment of groundwater levels—in wells or boreholes proximate to candidate impacted sites—provides supporting information. Specifically, a premise is that downward inflection of well water levels concurrent with (or approximately concurrent either prior to or soon after) declining surface water flows increases the likelihood of a construction impact to surface waters.

The approach used here is correlative in nature; it does not directly link changes in surface water flows to any particular source (the approach is not technically ―cause and effect‖). A basic premise, however, is that if water levels at wells proximate to the surface monitoring sites drop approximately concurrent with passage of the tunnel boring machine, then if surface flows decrease (compared to reference flows) concurrent with the well impact, then a reasonable presumption is that tunnel construction ―caused‖ the surface flow decreases.

Tunnel construction can affect groundwater levels which link to surface springs and streams. This linkage is most easily identifiable during baseflow conditions—typically in summer and early autumn when surface flow in the Mediterranean climate of southern California results solely from groundwater inputs. Consequently the best time to identify a potential construction impact to surface waters is during summer, rather than winter periods when rainfall and surface runoff can contribute to streamflow. An implication of this groundwater focus is that intermittent and ephemerally-flowing surface water is much less likely to be impacted by tunnel construction because naturally intermittent and particularly ephemerally flowing channels have limited or no groundwater sources. This means impact assessment for intermittent and ephemeral channels in the project area is generally very difficult because the preponderance of zero flow at these sites during both baseline and construction-period summers negates use of most quantitative impact assessment techniques, and potential construction impacts are masked in winter by precipitation/runoff processes.

In practice the primary assessment procedure (reference-impact approach) initially ocularly compares flows through time between one or more reference sites and a candidate impacted site during both the full spectrum (12-month) of flows and in more detail during a baseline period before any possibility of construction impact. If the ocular relationship between flows does not change during the construction period (compared to the baseline period), the candidate site is classified as not impacted. If a clear and drastic change occurs concurrent (or nearly so) with a downward water level inflection in a nearby well the site is tentatively classified as impacted. The 12-month comparison is followed by a comparison of flows solely during baseflow months (July through October inclusive) for both the anticipated pre- impact/baseline period and yearly thereafter. Figures 3 through 5 exemplify this procedure. In Figures 3 and 4 the full spectrum of flows at reference site 185 are plotted against the full spectrum of flows at candidate impacted site 56. At nearby well 911 water levels drastically declined starting in late 1998, with water level declines also evident slightly later in well 912. These water level declines were concurrent with tunnel mining in close proximity to the two wells. Flows at reference site 185 are uniformly lower than at candidate impacted site 56 during the baseline period prior to late 1998, but are followed in late 1998/early 1999 by greatly reduced flows at 56 compared to 185. Plotting of the 185-to- 56 baseflow (July through October) relationship for individual years from 1999 through 2011 (Figure 5) details the changed relationship in flows between the two sites and indicates a construction effect beginning between the summers of 1998 and 1999. If the post-baseline yearly flow pairings fall within the loci of the baseline flow pairings then no impact is indicated. For site 56, however, the post-baseline (1999-2011) pairings are below the range of baseline pairings (the blue triangles), indicating a construction effect. Figure 5 is somewhat atypical in that usually more baseline pairings are available than for site 56. The smaller the ―sample size‖ the less confident the results, but in Figure 5 the 1999-

2010 pairings are clearly outside the range of pre-1999/baseline pairings suggesting reasonable confidence that a construction effect exists.

If a clear and definitive ―impact‖ is not identified, both 12-month and baseflow-only comparisons are made between a second reference site and the candidate impacted site, with the intention that results of this second comparison will clarify the impact status of the candidate site.

The confidence level of the impact identification (objective 2) is a subjective determination. If the relationship between the baseline (pre-impact) flows at the reference and candidate impacted sites isn’t appreciably different from the relationship during the candidate impacted period, typically a ―no impact‖ determination is made. If evidence is conflicting a ―gray area‖ determination results.

The trend of impacted sites (objective 3) has three potential outcomes: (1) no evidence of rebound/recovery, (2) some evidence of flow rebound/recovery, and (3) recovery complete. The distinction is currently subjectively determined, although potentially a quantitative component could be added to this determination. If flows at the impacted site are starting to ―come back‖ particularly relative to flows at the reference site, then the impacted site is categorized as ―showing evidence of rebounding‖ (see for instance the 2006 to 2011 period at site 56 (Figure 4) where site 56 summer flows have increased (from 0 in prior, impacted years), but are not yet greater than concurrent flows at reference site 185. If flows at 56 were greater than at 185—as per the baseline pre-1999 period—then arguably flows at site 56 have recovered.

APPROACHES TO IMPACT AND RECOVERY ASSESSMENT

Reference-Impact

The reference-impact approach is exemplified in Figures 3 through 5 wherein a potential impact is identified at City Creek tributary site 56 in late 1998 when the pattern or relationship shifts between flows at the reference and candidate impacted site. This, as well as other, approaches have strong and weak aspects, and considerations for their usage.

A basic premise of the reference-impact approach is that the reference site(s) occurs in a hydro- geoclimatic regime similar as the candidate impacted site. This effectively means that the reference site(s) ideally is located close to (e.g., in the same catchment and at the approximately equivalent elevation) the candidate impacted site so that climatic, channel geomorphology, flow regime and magnitude (e.g., spring versus stream) and other variations between the sites are minimized.

Pros and cons—Benefits of the reference-impact approach include— Integration of confounding factors like climate change, natural meteorological and hydrological variability, drought, evapotranspiration, shallow sub-surface processes (e.g., bank storage), changes in channel morphology, and wildfire impacts that otherwise would need to be isolated and dealt with individually (i.e. well chosen reference sites should have climate, flow regimes and wildfire history similar or identical to candidate impacted sites). Flow differences, or similarities, between candidate impacted and reference sites are easily displayed and the technique is understandable by non-specialists. Potential differences in flow magnitude caused by anthropogenic activities can be well quantified.

Liabilities of the reference-impact approach include—

The potential inability to identify reference sites because candidate reference sites have been impacted. Candidate reference sites do not ―match‖ the characteristics of potentially impacted sites in having different flow regimes (surface water vs. groundwater dominated, perennial vs. intermittent/ephemeral), and/or differing catchment characteristics (small vs. large, or low-elevation rainfall-dominated vs. high elevation snow-dominated). Too few reference sites to provide confidence in the assessment. Subtle impacts may be less easily identifiable than abrupt impacts of the type illustrated in Figures 3 through 5. Different measurement times at reference and candidate impacted sites can add variability or ―noise‖ to the assessment. Flows in the lower main stem of Sand Canyon Creek, for instance, can vary over 40 gallons per minute (gpm) within a twenty-four hour period during some summer months. Measurements at two locations made a few hours apart may not be completely comparable simply because flows would be varying naturally over that time frame. However, if the measurement times are consistent through time at each site then the ―baseline‖ relationship should incorporate time differences between sites.

Although there are liabilities with the reference-impact approach, the usefulness of this approach comes down to the comparison of flows during a potentially impacted period versus a baseline period. If the baseline period has a relatively wide spread of ―points‖ (as might be generated by some of the liability considerations) but the location of the flows/points during the candidate impact period is still clearly outside the baseline spread, then a potential impact is identified. So even if the baseline spread is wide (i.e., variable and potentially with relatively few datapoints), an impact can be assigned if the post- baseline loci of flows is still outside the baseline spread.

The reference-impact approach is correlative; it does not identify specific causes for changed surface flow patterns. Although the coincidence in time of a well-defined tunnel mining impact (e.g., increased groundwater inflow at a specific tunnel location ―near‖ a surface monitoring site and/or well water decline associated with mining-induced groundwater inflows to the tunnel) with reduced flow at the surface is suggestive of mining as the cause for the surface flow decrease, this approach does not otherwise causatively link mining with surface flow changes.

Identification of reference sites—The reference-impact approach relies heavily on the flow history at a few reference sites. This reliance makes the choice and flow characteristics of the reference sites critically important. Ideally multiple reference sites would be available to strengthen any conclusions about impact status. The current assessment does include multiple reference sites when the ―signal‖ of an impact is questionable initially based on comparison to a single reference site.

A good reference should respond to factors like climate change and rapidity of recovery from wildfire in a manner similar to the candidate impacted site, and should have similar channel and riparian vegetation characteristics.

Attributes of a good reference site include— Spatial proximity to the candidate impacted site. Location within the same watershed, or less preferable, location in different, but same-sized watershed. Similar hydrologic regime to candidate impacted site. o Groundwater-sourced spring or rain-dominated stream. o Similar flow magnitude to the candidate impacted site. 17

o Perenniality (as an indicator of groundwater sourcing). o No evidence of construction impact (implying either distance from the tunnel alignment or good flow correlation with other candidate reference sites). o Concurrent (same day, hour) flow measurements with the candidate impacted site.

Specific reference sites—Unfortunately, many monitoring sites on the Arrowhead East (AE) alignment are candidate impacted sites. This means that relatively few AE sites are candidates for references. For AE four candidate references were initially identified, site 185 in lower Sand Canyon, site 44 in middle Little Sand Canyon, site 154 in the main stem of lower Borea Canyon, and the USGS gage located near Highland at the base of the City Creek canyon.

Flow at the USGS City Creek gaging station represents a much larger watershed than any watershed of the small, ―face‖ watersheds in the immediate ATP project area. Also the City Creek watershed extends much higher in elevation than any face watershed in the AE project area. These factors reduce the potential for the City Creek site to be a good reference, unless perhaps for the two Sand Canyon main stem sites, 117 and 103, that have relatively high flows, and the three ATP monitoring sites on City Creek itself (520, 625, and 151). Correlations between flows at the USGS gage and sites 117 and 103 are not bad, further implying some utility in considering the USGS gage as a reference for sites 117 and 103. The 91-year USGS City Creek flow record benefits from the availability of daily measurements that can be therefore synchronized in time with the weekly (or less frequent) measurements at project area monitoring locations. However, the close proximity of the City Creek tunnel to City Creek implies that flow at the City Creek USGS gage itself could be impacted by tunnel construction. Although this is relatively unlikely because most of the water passing the gage originates higher up in the watershed away from the tunnel, the potential for an impact at the GS gage can’t be entirely ruled out.

Site 185 is at the mouth of a tributary in the lower middle portion of Sand Canyon. It has recorded flows typically ranging up to 15 gpm (more usually 9 or 10 gpm winter maxima and 1-5 gpm during summer baseflow periods). Flow at 185 was generally perennial during the 1994-2011 period of record, but included a few 0 readings during the 2000-2003 dry period. Site 185 is a spring that does receive some winter rain-induced runoff so its hydrograph has both groundwater and surface runoff components. Flow at 185 correlates well with flows at other candidate reference sites (e.g., site 27 in Badger Canyon). Independent assessment by MWD staff and other FS staff show the usefulness in using 185 as a correlate for the other Sand Canyon sites. And those staff members agree that site 185 was not impacted by tunnel construction.

Site 44 is an adit with no upstream water source. It flows totally from sub-surface sources. The hydrograph of 44 matches closely the hydrograph of wells in the vicinity of 44 with very little within-year variation due to individual precipitation events. Between 1994 and early 2009 flow at site 44 peaked at approximately 17 gpm in 1998 and 2005 but otherwise has generally seen little variation superimposed on downward trend from 1999 to early 2005 and again from mid-2005 to late-2010. Flow at 44 correlates well with flows at other locations with groundwater-only sources. Unfortunately, there is a chance that 44 has been impacted by tunnel construction.

Borea Canyon site 154 receives water from a spring a few meters above the monitoring location. It also receives water from the entirety of Borea Canyon, including site 45. During the summer, much of the flow at 154 is presumably from the spring (e.g., groundwater sourced). However, some flow at 154 in the summer is probably contributed from up-channel. To the extent that site 45 may be impacted, potentially 154 is indirectly impacted also.

Sites 44 and 185 are not optimum for comparison to main-stem sites in any of the AE canyons because winter flows in particular at 44 and 185 can be considerably lower. This isn’t entirely ―bad‖ because winter conditions themselves are not optimum for identifying impacts; impact identification is best done by comparing summer baseflows that are entirely groundwater driven. Summer flows between sites 44 and 185 and the other sites on AE are more comparable. For this assessment site 185 is used as a reference for comparison to candidate impacted sites on the AE alignment.

Several monitoring sites are candidate references for the Arrowhead West (AW) alignment. These include sites 213, 214 and 27 (all in upper Badger Canyon), and 10, 11, 157 and 9 (all in upper Ben Canyon). These all have some surface runoff components to their hydrographs and all have generally low flows. For this assessment sites 11 and 27 are used as references for comparison to candidate impacted sites on the AW alignment.

The western portion of the AE alignment includes monitoring sites in the Strawberry and East Twin Creek catchments north of the alignment. Most of these sites are more distant from the AE alignment than the more southerly sites in the smaller Borea, Little Sand, and Sand Canyons. One site in particular, 645 is about 1 ½ mile from the alignment, the furthest of any site from either alignment. Site 645 is considered a reference site for applicable sites in the East Twin, Waterman, and Strawberry catchments.

Considerations—Several factors need to be explained to better interpret the scope and limitations of this analysis. Taken together these argue for a conservative interpretation of the available data and consequently a conservative approach to impact assessment.

Sources of variation in flow quantities—As discussed above, although the MWD spot flow measurements are generally made at the same time of day, diurnal flow change exists (and has been measured routinely to be in the 30-40 gpm range for lower Sand Canyon during summer periods when evapotranspiration is high during daylight hours). Some variation in flow measurement is anticipated to be associated with this dirunality. Nevertheless summer baseflow measurements are relatively consistent at many locations on a monthly timeframe (Tom Hibner personal communication February 12, 2009).

Other sources of variation in the flow data are: Measurements made on differing days—Measurements at some reference sites were not made on the same day as measurements at candidate impacted sites. This introduces possible temporal sources of variability. To reduce this ―error‖ flow comparisons are restricted to measurements made only a few days apart (typically 3 or less days). Although this is not an ideal solution Hibner’s aforementioned analysis suggests that during baseflow periods—the period of most interest--flows are relatively consistent at any given monitoring site at the same time of day. Observer ―error‖, which may vary by flow magnitude (e.g., greater absolute error at higher flow magnitudes and greater percent error at lower magnitudes). In particular measurement variations at Sand Canyon site 103 seem to be more variable than anticipated, and are relatively more variable than measurements at the next upstream site, 117. Variable site conditions--Measurements are best made at bedrock channel locations where there’s no opportunity for water to flow in sediment (underflow); sedimentation—as produced by fire-flood events--can alter channel morphology so that measurement locations may not remain sediment-free. Measurement ―error‖, produced when differing measurement techniques are used (e.g., volumetric measurement timing a known amount of flow vs. measurement with a flow meter).

Groundwater dynamics--Potentially intricate and largely unknown groundwater pathways influence the timing and amount of well water level changes and the interaction of those changes with surface water 19

sources (e.g., springs and seeps). An added complexity is the dynamic nature of these relationships in time and space.

2003 wildfire—The Old wildfire in autumn 2003 affected all parts of the project area. From a hydrologic perspective, removing massive amounts of vegetation changed erosion and evapotranspiration rates. The erosion changes altered channel morphology (e.g., sedimentation and erosion) and temporarily filled several channels. This complicated surface water monitoring and potentially added variability to the flow dataset. Reductions in evapotranspiration changed the amount of water entering the channels as the ―left over‖ after plant demands were met; after the fire theoretically flows should have increased due to the abrupt reduction in water use by vegetation. Neither of these factors is readily quantifiable. Hydrologic effects of the Old fire also vary with time since the fire. Evapotranspiration itself changes through time as vegetation returns. Evapotranspiration ―recovery‖ is occurring and for at least some aspects of hillslope brush vegetation is approximately 99% complete by 9 years after the fire (Lave and Burbank 2004).

Comparison to Flows at USGS Gages

Daily flow magnitude data are available from four USGS stream gages that span the ATP area: Devil Canyon Creek near San Bernardino, City Creek near Highland, East Twin Creek near Arrowhead Springs, and Plunge Creek near East Highland. Summary attributes of the watersheds these gages monitor and the available data are given below:

Gage Devil Cyn City E Twin Plunge Record Duration 1920-2012 1919-2012 1919-2012 1919-2012 Elevation 2080’ 1580’ 1590’ 1590’ Drainage Area 5.49 sq mi 19.6 sq mi 8.8 sq mi 16.9 sq mi

These gages are located about ¼ mile west of the west portal of the west tunnel (Devil Canyon gage), between the west and east tunnels (East Twin), about ¼ mile east of the east portal of the east tunnel (City), and about 3 ½ miles east of the east portal of the east tunnel (Plunge). These gages are all at the base of watersheds that are appreciably larger than the largest face watershed within the ATP, Sand Canyon. The flow records for these gages span more than 90 years and therefore allow current and recent flows in the ATP area to be qualitatively compared to a long-term perspective. For instance, if flow at the GS gages is ―above average‖ or greater (defined as greater than the 76th percentile of the long-term flow for the given date—or the 7, 14, or 28 days prior to the given date) then intuitively flows at the MWD sites should be expected to be greater than average. Because of their larger watershed contributing areas the USGS gages should represent environmental conditions (e.g., climate, evapotranspiration) on a broad spatial scale that includes the ATP area. If GS gages are above average and the MWD sites are below average, presumably something has caused the low flow at the MWD sites. Where applicable, data from these gages are used as references for main channel MWD monitoring sites in City and Devil Canyon Creeks.

SITES ASSESSED FOR POTENTIAL CONSTRUCTION IMPACT

Since 2000 a primary, ongoing ATP activity has been identification of impacted stream and spring sites. Although this present report is the first formal impact assessment, it benefits from years of informal analysis and discussion among staff of MWD, the San Manuel tribe, and the FS. For a variety of reasons this current report does not assess impact potential at all sites, and text describing the reason(s) for non-

assessment is provided for each site. Typically sites are not assessed because they are intermittent or ephemeral waters that have zero flow in summer when tunnel-induced effects on groundwater would be most evident. Some other sites were established too late to provide an adequate baseline prior to potential impact.

RESULTS

Impact identification and current status are summarized below, followed by explanation of these results on a site-by-site basis.

Definitely directly impacted: City o 56, 181, 58 Sand o 48, 636, 53, 54 Little Sand o 510, 44 Borea o 45, 154 (possible upstream contribution to impact) Waterman/Sycamore o 65 Sycamore o 156

Definitely indirectly impacted (upstream impacts cause flows at these sites to be less than natural): Sand o 117 & 103 Little Sand o 509, 155 Waterman/Sycamore o 17

Grey for direct impact: Sand o 635 Strawberry/East Twin/Waterman o 189, 642

No evidence for impact: Coldfoot o 210, 55, 209 Sand o 51 Strawberry/East Twin/Waterman o 93, 94, 38, 134, 624, 628, 644, 191, 645, 629 Sycamore

o 205, 20, 30 Badger o 27, 213, 214 Ben o 9, 10, 157, 11 Devil o 153, 110

Cannot assess: City o 59, 60, 151, 515, 520, 622, 625, 630, 631, 632 (Measurements at all of these sites began too late (1997) to develop adequate baseline for comparison to impacted flows that could have begun in 1998. Also many of these sites are often dry in summer.) Little Sand o 637 (Insufficient summer/baseline period non-zero flows to develop a baseline flow relationship.) Strawberry o 120, 676, 677, 678, 679 (Baseline data available only from 2000 to 2003, a very dry period, with almost exclusively zero summer flow, therefore baseline relationship would not include complete range of flow to allow comparison to potential impact period flows.) East Twin o 90 (Concrete-capped well. Six-month monitoring frequency too infrequent to allow impact analysis.) o 201 (Per Tom Hibner (MWD): Covered by a debris flow after the 2003 fires and subsequent rains. The site ended up being under 5+ feet of sediment, therefore, flow likely occurred underneath the sediment.) Waterman o 93 (Ephemeral, insufficient summer non-zero flow to generate baseline relationship.) Sycamore o 95, 182 (Water artificially diverted so natural flow magnitude unknown.) o 627 (Ephemeral, cannot develop baseline relationship.) Badger o 28 (Ephemeral, cannot develop baseline relationship.) o 26 (Data available only 1994-2003, prior to any potential tunnel-related impact.) o 21 (Data available only 1994-2001, prior to any potential tunnel-related impact.) o 152 (Prior to mid-2006 measurements made every 6 months, too infrequent to develop baseline relationship.) Devil o 8 (Ephemeral, insufficient summer non-zero flow to generate baseline relationship) o 190, 193, 620 (All in main channel of Devil Canyon Creek with flows too high to confidently use reference site within immediate project area. Correlation with USGS Lytle Creek gage as reference is too poor to allow impact assessment.)

Site-by-site Specifics

Detailed explanations of the rationale for impact presence, confidence of the impact identification, approximate date of impact, and current impact status are ordered from the most eastern/lowest elevation

locations to the most western/highest elevation sites (from City Creek tributary cluster 56/58/181 to upper Devil Canyon site 193).

City Creek

City Creek, along with Devil Canyon, Waterman, East Twin, and Strawberry Creeks, drains relatively large portions of the San Bernardino Front Country. Elevations in these watersheds range up to the top of the San Bernardino Mountains plateau. In contrast, other monitored locations in the ATP area are in smaller, lower elevation ―face‖ watersheds (―Coldfoot‖, Sand, Little Sand, Borea, Sycamore, Badger, and Ben Canyons). Consequently the geology, climate, and hydrologic regimes of City, Devil, Waterman, E. Twin, and Strawberry Creek watersheds can differ from those of the smaller catchments. For instance, the larger watersheds annually receive precipitation in the form of snow, a very rare event in the lower catchments. These differences add variation to the surface water regimes in the project area, one major difference being much greater flow magnitude in the larger watersheds than in the smaller ones. The differences in flow regimes drive other differences, for instance different habitat opportunities for biota in the project area.

Surface water monitoring locations in the City Creek catchment include three perennial, high-flow sites on the main stem of City Creek (151, 625, and 520) plus nine sites on four west-side tributaries. All of these sites are east of the tunnel alignment. Most of the tributary sites flow intermittently or ephemerally. Three sites in ―Coldfoot‖ Canyon, the highest elevation tributary to City Creek in the ATP area, are addressed separately below.

Four monitoring wells are potentially relevant to surface flows in the City Creek catchment: moving from north to south wells 911, 912, 199, and 913. Well 911 is west of the alignment, about 0.3 mile, and is about 0.75 mile from a six-site cluster of monitoring sites (60, 59, 630, 631, 632). 911 started a construction-induced decline in water level late in 1998. Well 912, a few hundred feet from the alignment, is about 0.3 mile west of the site 56/181/58 cluster. A downturn in water levels in late September 1998 for 912.1 and a similar downturn in mid-August 1998 for 912.2 indicate a construction impact. Wells 199 and 913 are located a few hundred feet from each other and both are very close to the alignment. Both of these wells are also close to the east portal of the east tunnel, with 199 about one- quarter mile distant and 913 even closer. Well 199 may have been impacted by construction activities no later than mid-August 1998, when water levels were steeply declining, but the absence of data before mid-August 1998 precludes a firm impact assessment for this well. A construction-related impact at well 913 began in mid-July 1998. Two other wells, 959 and 957, are 0.5 and 0.7 mile west of the alignment respectively, located within a few hundred feet of the border of the San Manuel Tribe reservation. There is no evidence for a construction impact at either of these wells.

Site 56 — This spring site is located partway up a drainage tributary to lower City Creek. Prior to tunnel construction the baseline record indicates that flow at site 56 was perennial. The channel above site 56 flows intermittently so that site 56’s hydrograph combines characteristics of a spring—with relatively constant summer baseflow—plus winter flow spikes associated with runoff-producing rainfall events that wet the channel above 56. Although the record is relatively short, prior to late 1998 flow at 56 was seldom below 6 gpm (Figures 3 and 4).

Flow at site 56 declined rapidly starting in late 1998 in comparison to flow at reference site 185. Late 1998 coincided with a decline in water levels at well 911 and slightly later than the construction impact began at well 912. Construction-related inflows to the City Creek tunnel began rising dramatically in

October 1998 and maxed out near 1800 gpm in February 1999. Mitigation water was applied immediately above site 56 during each summer from 1999 to 2003 and sometimes in fall and early- winter, typically at a rate of 1-2 gpm.

Comparison of summer baseflows (Figure 5), in temporal conjunction with known construction effects at nearby wells, suggests a construction impact from 1999 through 2011 when the 56-185 baseflow relationship was way below the pre-1999 regression line and outside the range of the pre-1999 relationship. Two secondary factors are relevant to interpretation of this figure: 1) irrigation was is included in the plotted flows at site 56 for 1999 through 2003, and 2) the number of pre-1999 flow pairings is relatively small (i.e. a small sample size). The irrigation addition actually pushes the 56 flow values up, toward the pre-1999 points, and argues for an even greater construction effect than shown on the graph for 1999 through 2003. The large separation between the pre-1999 and 1999-2011 points argues strongly that the small pre-1999 sample size does not negate the conclusion of an impact from 1999 through 2011.

The dramatic nature of the late 1998 flow decline, in combination with a concurrent lowering of well water levels and significant tunnel inflows, implies a high likelihood of construction impact. Although natural flows resumed early in 2005, as of September 2012 the pre-impact flow relationship with site 185 (e.g., 56 flows typically higher than 185 flows) had not resumed on a consistent basis.

Site 181 —Stream site 181 is located approximately one-quarter mile downstream from site 56 and about 800 ft east of the tunnel alignment. An ephemeral channel joins the 56 channel above 181. There is also probably a minor seep or spring water source upstream a few hundred feet from 181. Site 181’s hydrograph includes winter flow spikes indicting that flow at 181 is influenced by surface and near- surface flow paths (181 is a typical Front Country stream system, incorporating both upstream spring water as summer baseflow plus a winter rain water source). Although the record is relatively short, prior to late 1998 flow at 181 was seldom zero and typically greater than flow at reference site 185 (Figures 6 and 7).

Similar to flow at site 56, flow at 181 declined drastically during winter 1998-1999 and remained at zero for most of the period from spring 1999 to autumn 2004. During much of this period mitigation water was applied at site 56 although the 1-2 gpm application probably was taken up by evapotranspiration and seldom reached the 181 monitoring location. The dramatic nature of the late 1998 flow decline, in combination with a concurrent lowering of well water levels, and significant tunnel inflows, implies a high likelihood of construction impact. Natural flows resumed early in 2005 but as of October 2011 the pre-impact flow relationship with site 185 (e.g., 181 flows typically higher than 185 flows) had not resumed consistently nor had the summer baseflow relationship between sites 185 and 181 returned to the baseline (pre-1999) resumed (Figure 8).

Site 58—Flow is measured at stream site 58 at the base of a waterfall that typically has exhibited minimal flow (less than 0.5 gpm) except for sporadic rain-induced higher winter flows. This measurement site is about one-quarter mile from City Creek, about one-quarter mile from site 56, and about one-third mile from the tunnel alignment. Prior to tunnel construction flow at site 58 was perennial with flows seldom recorded below 0.05 gpm.

Although construction of the City Creek tunnel segment clearly de-watered nearby sites 56 and 181, site 58 was not thought to have been impacted and the time series comparison of flows at sites 58 and 185 does not show an obvious impact (Figures 9 and 10). Comparison of summer baseflows at 58 and reference site 185 using measurements made within three days, however, suggests otherwise, with many baseflows at 58 starting in 1999 being lower than pre-1999 flows (in comparison to flows at 185) (Figure 24

11). The baseflow comparisons suggest an impact at site 58 starting some time in 1999, approximately concurrent with water level drops in nearby wells 912 and 911. Also since construction began zero flows have been recorded at 58 in the summers of 2000, 2003, and 2005-2011 (there were no zero flows prior to the beginning of construction). Rebound appears to have started in 2001 but is not complete to the extent that flows at 58 are still (through September 2011) usually lower compared to flows at 185 than the pre- impact relationship would dictate. One caveat to the determination that site 58 is impacted is the relatively small baseline sample size. Nevertheless the common and continuing occurrence of zero flow at site 58 through summer 2011 strengthens the call that this site was impacted and that recovery is not complete.

Site 59—Located about 0.6 mile east of the alignment, and equidistant (approximately 0.95 mile) from wells 911 and 912, flow at this site each summer from 1999 through the end of record in 2010 was zero. The relatively sporadic data before summer 1999 shows some of the pre-1999 summers to have measureable flow (Figures 12 and 13). A visual change in flow pattern between site 59 and 185 before and after 1999 is not however, well identified; flows at 185 were generally above those at 59 from 1995 through 2010. Unfortunately, measurements made at site 59 and reference site 185 were seldom synchronous enough to allow confident comparison of flows between the sites so that the reference- impact comparison technique during the summer baseflow period isn’t applicable. Consequently the impact status of site 59 is undeterminable with the techniques used.

Site 60—This site, characterized by MWD as a spring, is about 0.4 mile east of the tunnel alignment and approximately equidistant from wells 911 and 912. These wells registered construction impacts initially during the last 4 ½ months of 1998. Between 1995 and 2011 non-zero measured flows occurred only between September 1997 and February 1999 at site 60, in magnitudes ranging up to 6 gpm. Because no flows were measured after February 1999 the comparison techniques used at other sites are not applicable. Although a dive in flow magnitudes in late 1998 and early 1999, culminating with the final measurable flow observation in February 1999, is roughly concurrent with the timing of construction-related impacts at wells 912 and 911, attributing the flow decline to tunnel impacts can’t be comprehensively assessed. For this site impact status is undeterminable.

Site 515—Located about 0.6 mile east of the alignment, this site is about 0.75 mile from well 912 and slightly further away from well 911. Impacts to well 912 began in mid-August 1998. Prior to mid- August 1998 less than one dozen summer baseline measurements are available from site 515. This lack of baseline data precludes use of the baseline comparison technique. The few candidate baseline measurements were not concurrent with measurements at reference site 185, precluding the reference- impact comparison technique, and no other satisfactory reference site is available. Visual comparison of flows before and after the candidate autumn 1998 impact date is inconclusive. Some late 1998 and early 1999 flows at 515 were higher than those at 185, others lower. And although after mid-1999 flows were always lower at 515, the impact status of site 515 is undeterminable.

Site 632—This site is located within 100 ft of site 515, and consequently would be expected to be impacted (if at all) during the later part of 1998, or later. Unfortunately, measurements at site 632 began in June 1998, too late to provide a sufficiently long time period before potential impact. Consequently, the impact status of site 632 is undeterminable.

Site 630— Located within 100 ft of site 515, this site would consequently be expected to be impacted (if at all) during the later part of 1998, or later. Unfortunately, measurements at site 630 began in June 1998, too late to provide a sufficiently long time period before potential impact. Consequently, the impact status of site 630 is undeterminable.

Site 631—As part of the site 60/515/59/630/632/631 cluster this site is closest to the main stem of City Creek and 0.75 mile from the closest well, 912. As with most other sites in this cluster, measurements at 631 began in 1998, providing an insufficient time period for generating a potential pre- impact baseline. Consequently, the impact status of site 631 is undeterminable.

Site 622—Located a quarter-mile from the base of the mountain front, site 622 is not within the City Creek catchment. Rather it’s positioned within a ―face‖ catchment between Sand Canyon and the City Creek watershed. Because site 622 is located closer to City Creek than Sand Canyon, it’s grouped here with the City Creek sites. Site 622 is 0.6 mile due west of well 913 and about 0.5 mile southwest of well 912. Based on the construction impact dates for these wells, impact at 622 could be expected in mid- 1998, or later (if at all). Monthly monitoring began at site 622 in January 1997, with four summer measurements in 1997. As with surface sites in the 60/515/59/630/632/631 cluster the monitoring period prior to the potential earliest impact date is too short to generate an adequate pre-impact baseline. Consequently, the impact status of site 622 is undeterminable.

Site 151, 520, and 625—These three sites are all on the main stem of City Creek, with 151 the northernmost site located about 1.1 mi east of the alignment, 520 the southernmost located approximately 0.4 mi almost due south of the east portal of the east tunnel, and 625 about 0.2 mi east of both the eastern tunnel portal and well 913. At all three sites only two summer measurements were made prior to summer 1998, when impacts initially could be expected at sites 520 and 625. Consequently measurements are insufficient at these two sites to develop an adequate baseline relationship. The timing for a potential impact at site 151 is problematic in that conceivably it could have been impacted (it at all) during either, or both, the early City Creek mining and the much later second mining phase. If an impact at site 151 occurred during the City Creek mining, too few measurements are available to generate a baseline relationship. If an impact at site 151 occurred during the second mining phase, unfortunately concurrent measurements at site 151 and reference site 185 are too rare to allow development of a baseline relationship. Attempts to use the USGS Plunge Creek gage data as a reference for site 151 were not successful. Consequently any impact at all three of these sites is undeterminable with the assessment approach used here.

“Coldfoot” Canyon

Coldfoot Canyon is a 273-acre drainage with two intermittently-flowing tributaries that join toward the base of the canyon. This catchment drains from McKinley Mountain east to City Creek. Two surface water monitoring sites are positioned on the northern tributary (210 and 55), with a third site (209) just downstream of the confluence of the two tributaries. All three locations had numerous zero-flow measurements during most summers of record. Well 956, located a few hundred feet west of the tunnel alignment, is the closest well to all three Coldfoot surface sites—0.5 mile west of 210 up to 0.9 mile west of 209. Water levels in both intervals of well 956 began to turn down in late-February 2007 followed by a steeper decline beginning in mid-May 2007. Well 959 is more distant from Coldfoot, over one mile away from the closest surface site, 210. Construction impacts began at well 959 in late summer 2007.

Site 210—Of the three surface water monitoring sites in Coldfoot, site 210 is closest to the tunnel alignment, about 0.4 mile distant. There’s no evidence of a construction-related effect from ocular comparison of flows at 210 and reference site 185 (Figures 14 and 15); flows at 185 are typically greater than those at 210 except for sporadic, runoff-induced spike flows at 210 in winter and spring. Presuming a potential construction impact sometime in early 2007 (per well 956), comparison of baseflow measurements at 210 and 185 before and after 2007 show no evidence for a construction impact (Figure 16). Because many baseflow measurements at 210 were zero, the pre-2007 baseflow relationship has

relatively few non-0 flows (for 210). Nevertheless, the post-2007 comparison places all points well within the pre-2007 range.

Site 55—Located a few hundred feet downstream from site 210, site 55 is about 0.45 miles east of the tunnel alignment. It’s hydrograph is similar to 210’s, with many zero measurements in summer. Also similar to 210, there’s no ocular evidence for a construction-related impact when comparing 12-month flows at 210 with those at 185 (Figures 17 and 18). Baseflow period comparison (between sites 55 and 185) before and after 2007 shows no evidence for a construction impact at 55. Again, zero flow dominates the baseflow comparison but the post-2007 flows are well within the range of the pre-2007 values (Figure 19).

Site 209—At 0.8 mile from the alignment, site 209 is the furthest of the three Coldfoot Canyon sites from the tunnel. During most summers flow at this location has been 0. There’s no visual evidence for a construction-related flow impact at site 209. Flow at site 185 is usually greater than flow at site 209 during summer baseflow periods both before and after the anticipated 2007 potential impact date (Figures 20 and 21). And baseflows at 209 were actually greater between 2007 and 2010 than would have been expected from the pre-2007 relationship (Figure 22). On this basis there’s no reason to judge site 209 as being impacted by tunnel construction.

Sand Canyon

Sand Canyon, the largest of the lower Front Country drainages within the ATP area, occupies about 2,040 acres west of the lower portion of the City Creek catchment (Figure 1b). During the MWD surface water monitoring period, for most years the lower main stem of Sand Canyon flowed perennially, although in 2002 and 2003 there were no surface flows during some summer months in the main channel. Several tributary channels join the main stem throughout its length down to the boundary of Forest Service and tribal lands. The larger of these tributaries typically flow perennially at their junctions with the main stem. Snow and below-freezing air temperatures occurred rarely in Sand Canyon during the monitoring period of the ATP. Biological surveys in 2003 identified sections of the main channel to be habitat for Least Bell’s Vireos and Southwest Willow Flycatcher.

Nine surface water locations are monitored by MWD staff, most recently at weekly intervals (Figure 1b). Three MWD monitoring sites are north of, but very close to, the tunnel alignment. Site 48 is on a small tributary receiving water as a spring that is augmented in winter and spring by up-channel surface runoff. Site 635 is on the main stem just above the site 48 tributary. This location flows continuously during winter and spring months but typically dries up during the summer. Site 636 is at the base of the northern-most major tributary in Sand Canyon. Prior to tunnel construction it flowed perennially, including the 2000-through-2004 dry period. These three sites are all less than one-fifth mile from the tunnel alignment, with site 636 being almost on top of the alignment. Sites 53 and 54 are at the base of tributaries entering the main channel from the west and east respectively. Sites 51 and 185 are at the base of tributaries, again entering from the west and east respectively, further down-canyon. The last two surface water monitoring sites, 117 and 103, are on the main stem, with 103 located a few hundred feet above the property boundary with the San Manuel reservation. Distances from the mid- and lower- canyon sites to the alignment range from about 0.4 mile (sites 53 and 54), to 0.95 mile (site 103). During some portions of the project both MWD and tribal staff monitored main stem flow in the lower portions of the main channel at 10-to-15 minute intervals. The tribe’s H-flume was immediately south of the reservation’s northern border and MWD’s V-notch weir was located near manually-monitored site 117.

In addition to the nine surface water monitoring sites, six terraces were identified as vegetation locations of importance in the middle and upper main channel of Sand Canyon. All six terraces were irrigated in 2007 and the lower three were again irrigated, at a lower rate, in 2008. The upper canyon monitoring sites were also mitigated, for varying periods and magnitudes from 2007 through 2011.

Four monitoring wells are located within the NFS portion of Sand Canyon, 956, 955, 910, and 959. Of these 956 is by far the closest to the alignment, about 0.1 mile distant. It was initially impacted by tunnel construction in February 2007 (see also the City Creek text above). 956 is between 0.5 and 0.6 miles from upper canyon sites 48, 635, 636, 53, and 54, and more distance from lower canyon sites 51, 185, 117, and 103. Well 959 is located on the eastern slope of the canyon, about 0.5 mile west of the tunnel alignment. It ranges from 0.4 to 0.6 mile from the middle and lower canyon surface monitoring sites and was probably impacted September or October of 2007. The final two Sand Canyon wells, 955 and 910, are in or very near the main stem channel. 910 is located in lower Sand Canyon between surface sites 117 and 185. There is no evidence of a construction impact at well 910. 955 is located within a few hundred feet of both the tunnel alignment and surface site 636. 955 is further from the other upper canyon surface sites but still within one-quarter mile of sites 48, 635, 53, and 54. The upper interval of well 955 experienced a construction-related impact in early October 2006, preceded by an impact to the lower interval sustained two weeks earlier.

Wells 909 and 958 in L Sand Canyon are moderately close to some of the Sand Canyon surface monitoring locations and could potentially provide information on dating of possible construction impacts for some of the Sand Canyon sites. Well 958 is about 0.5 mile west of Sand site 103 and about 0.6 mile from site 117. Well 909 is about 0.7 mile from Sand sites 53 and 51 and another 0.1 mile more distant from Sand sites 54 and 185. Well 909 was impact initially in late-December 2005. There’s no obvious evidence of construction-related impact to either interval of well 958.

Site 48--Spring site 48 is located approximately 50 ft above the main stem of Sand Ck on a steep west-facing slope. This location is about 0.15 mile from the tunnel alignment. Although a small catchment exists upslope of site 48, prior to tunnel construction flows at 48 generally followed a ―classical‖ spring/ground-water dominated hydrograph paralleling closely water levels at proximate wells (flows at 48 increased during high groundwater recharge periods—1998 and 2005—and otherwise showed little influence of winter rains). Even during the ―extreme‖ dry period from 2000 to 2004 summer baseflows at 48 were in the 3-5 gpm range (Figure 23 and 24).

Data irregularities from 2007 on complicate impact assessment at this site. In mid-May 2007 irrigation began at a point several feet above the site 48 monitoring location and continued through 1/11/08. Thereafter irrigation occurred continuously from the springs of 2008, 2009, and 2010 through the early winter months of each year, except 2010, when irrigation ceased in late October, and 2011, when irrigation water was applied for two weeks in July and from August 30 through October 4. The measurements at 48 include any mitigation water. Mitigation water measurements were not always made directly at the irrigation pipe itself but were sometimes derived as the difference between ostensibly known irrigation application rates elsewhere. Determining natural flow magnitude after mitigation began requires subtracting the irrigation rate from the measured flows at site 48. This differencing procedure had numerous anomalies, for instance in summer 2007 over one-half of the differenced values were negative, a clear physical impossibility, with the negative values ranging to above -5 gpm. On one summer 2008 date zero flow was reported at site 48 but the irrigation rate was reported as 18 gpm. These anomalies lower the reliability of any numerical assessment of impact at site 48.

Through March 2007, for most of the period of record, flows at 48 were above those at control site 185 (Figures 23 and 24). An unprecedented, precipitous drop in flows at 48 in spring 2007, to near-0 levels 28

by mid-May, in combination with stable flows at site 185, indicates a potential construction impact (Figures 24). This interpretation is supported by the early 2007 abrupt drop in water levels at well 956.1 (about ½ mile from site 48), presumably related to tunnel construction. The steepness of the site 48 flow decrease in spring 2007, coupled with the lack of return (through October 2011) to the pre-impact flow relationship with site 185, indicates a high likelihood that 48 has been impacted. Summer baseflows in parts of 2007, and much of 2008-2011, are below those of the pre-2007 baseline period (Figure 25), further supporting the existence of an impact at site 48. Some natural flow has resumed during summer months, and the flow regime at site 48 has marginally returned to the pre-impact pattern (as of September 2012).

Site 635—Monitoring site 635 is the upper-most surface water site in Sand Canyon. It is located on the main channel within a few dozen yards of spring site 48. Flows at 635 have been zero during many summers since routine monitoring began in November 1998 (Figure 26. Note logarithmic y-axis used because the widely differing magnitudes of flows at sites 635 and 185 can be better represented logarithmically than linearly).

Evidence for a construction impact at site 635 is somewhat conflicting. Water levels at well 956.1 (located about ½ mile from 635) were impacted starting in February 2007, suggesting that any impact at 635 could be expected to start late 2006 through spring 2007. Summer flows at both 635 and reference site 185 dropped to zero (or near-zero) during the 2000-2003 dry period (and before any anticipated impact). This relationship changed in summer 2007 (and thereafter through 2010) with summer flows at 185 becoming greater than those at 635 (Figure 26). In particular, 185 had measureable—but low--flows during all summer months of 2007 through 2010 concurrent with zero flows at 635 in July and August for each of these years. This change in flow relationship between the two monitoring sites, concurrent with the construction-related impact at well 956, suggests an impact at 635 at or before summer 2007.

However, a more detailed look at summer baseflows does not support a changed flow relationship between the two sites (Figures 27 and 28). Summer 2007-2010 flows fall within the range of the pre- 2007 relationship. A preponderance of zero summer flows at 635 from 2007 through 2010 weaken the usefulness of the baseflow comparison technique. On balance, the existence of a construction impact at site 635 is uncertain. 2011 summer flows at site 635 clearly show a rebound to well within the range of the baseline (pre-2007) flow relationship.

Site 636—At the base of a major tributary to upper Sand Canyon, site 636 is located almost on top of the tunnel alignment and within a few hundred feet of well 955. Flows at 636 were perennial from the beginning of routine monitoring in late-summer 1998 until summer 2007. This site was irrigated from May or June until December or January of 2007 through 2009, July through October 2010, and late- August through early-October 2011. Impact analysis for this site is complicated, as per site 48, by the irrigation. Natural flows were calculated by subtracting irrigation flow from the measured flows. Many ―calculated‖ natural flows during irrigation were less than zero. These were set to zero for this impact assessment.

The relationship between flows at sites 636 and 185 (Figure 29) changed in 2007. Prior to late spring 2007 flows at 185 were always less than those at 636. During the summers of 2007 through 2010 this trend reversed, and natural flow at 636 during much of this period was consistently zero. The changing relationship in flow between the two sites, concurrent with construction impacts at wells 956 and 955.2, suggest a construction impact at 636 from spring 2007 to at least late summer 2010.

Detailed baseflow assessment (Figures 30 and 30a) comparing same day, July through October measurements at the two sites confirms a construction impact from 2007 through much of 2010. Rebound has marginally returned to the pre-impact pattern as of September 2012.

Site 54 is 25-50 yards up a tributary on the east side of upper Sand Canyon (Figure 1b). Flow at 54 before tunnel construction was intermittent-to-perennial—perennial during and after wet years (e.g., 1998-2001) but dropping to zero during the summers of drier years (e.g., 1996-1997, 2004). Spikes in flow during most winters imply this site is influenced by surface runoff, and the zero flows in many summers suggest relatively little inputs from groundwater sources.

Impact to well 956, the nearest to site 54, began in February 2007. Irrigation 200-300 yards above site 54 occurred during the summers of 2007 through 2010 and extended into the early winter of 2008. Consequently the impact analysis for this site is less robust than for non-irrigated sites because of the added uncertainty on actual irrigation amounts, the amount of evapotransporation occurring between the irrigation outflow point and the flow measurement location, and the consequent determination of proxy natural flow rates. Visual comparison of natural flows (irrigation water subtracted out) at sites 185 and 54 does not show a pronounced change in pattern in the 2007-2008 period versus earlier monitoring; summer flows at 54 were lower than those at 185 during portions of the pre-2007 monitoring (Figure 31). Summer baseflow comparison also does not identify a strong case for a construction-related impact at site 54; natural flows for most summers are within the range of the baseline, pre-2007 period (Figure 32). However, the calculated zero flows at site 54 through all of summer 2008 and most of 2007 are generally outside the pre-2007 baseline range (Figure 33). On balance, there is moderate evidence for a construction effect at site 54 during the summers of 2007 and in particular 2008. 2010 and 2011 flows at site 54 clearly show a rebound to within the range of the baseline (pre-2007) flow relationship.

Site 53 monitors flow from a tributary entering middle/upper Sand Canyon from the west (Figure 1b). Wells 956 and 910 are about equi-distant from site 53, although the closest well, 955, is about one- quarter mile due north of 53. The site itself is about 0.4 mile south of the alignment. Before tunnel construction flow at 53 was perennial, although during particularly dry periods (e.g., 1996, 1997, 2001 through 2004) summer flow was less than 1 gpm. Spikes in flow during most winters imply this site is influenced by surface runoff, and the near-zero flows in many summers suggest relatively little inputs from groundwater sources.

Well 955 was impacted initially in October 2006. Irrigation at site 53 occurred from late spring through early winter in 2007, 2008, and 2009, from mid-June to late-October in 2010, and the entire month of September, 2011. As with most other Sand Canyon sites, the impact analysis for site 53 has more ―noise‖ because of the need to back-calculate natural flow amounts. Visual comparison of natural flows (irrigation water subtracted out) at sites 185 and 53 show a change in pattern starting in August 2006, when summer flows at 53 dip below those at 185 (Figures 34 and 35). This pattern persists through the end of record in summer 2011. Summer baseflow comparisons (Figures 36 and 37) further document a construction impact at site 53, particularly during the summers of 2007 through 2009. Flows at 53 started to rebound during summer 2010 with continuation through September 2012.

Site 51 monitors a small tributary entering the main stem of Sand Canyon Creek from the west, in the lower mid portion of the canyon (Figure 1b). As with sites 185, 117 and 103, this site is approximately equidistant from various locations on the tunnel alignment--ranging from 0.5 mile due south of the alignment up to 0.7 miles due east. This is relevant because potentially these sites could be impacted as early as 2000, when mining ended in the City Creek tunnel section, and between September 2006 (impact at well 955.2) and 2008 during the later stages of mining on the east alignment of the Arrowhead tunnel. Well 910, the closest well to site 51 was not, however, impacted. 30

Flow at site 51 is ephemeral; measureable flow occurred primarily during the winter and spring months of 1997-2000 and 2004-2005 (Figure 38) (measurements ceased 3/31/09). After spring 2001 only seven non-zero measurements were made, too few to conclusively use the reference-impact technique. The preponderance of zero flows at site 51 after July 2000 complicate the assessment of potential impact at 51. Considering baseflows only does not identify a potential construction impact; 2007 and 2008 flows are not outside the loci of pre-2007 points (Figure 39) and the common occurrence of zero flows prior to September 1997also suggests no impact. The likelihood for impact to site 51 in 2000 is difficult to assess because a severe dry period began about this time, reducing flows at both sites 185 and 51, and therefore potentially confounding the assessment by driving flows to zero at site 51 (as they had been during dry periods in the mid-1990s (Figure 38). Although a decision on impact potential is not well- supported because of the preponderance of zero flows, on balance it appears that no construction-related impact occurred.

Site 117—At its closest point to the tunnel alignment, site 117 is about 0.75 mile south. It is also within one mile of the eastern leg of the alignment at McKinley Mountain. Consequently 117 could conceivably have been affected by mining during both the City Creek 1999 activities and much later as mining occurred along the east-west alignment. Because of its location in the main channel of lower Sand Canyon any impact at site 117 is probably indirect as a result of impacts to upstream locations closer to the tunnel alignment.

During the monitoring period flow at site 117 was perennial except for brief periods in the summers of 2002 and 2003 (Figure 40). The hydrologic regime is definitely that of a stream, with rain-induced flow spikes in winter and groundwater-driven flow in summer. As with several other Sand Canyon monitoring sites, assessment of any construction impact at 117 is complicated by the contribution of irrigation water from upstream sites plus unquantifiable flow ―losses‖ due to evapotranspiration. Irrigation upstream of 117 occurred periodically from 2007 to 2011. Subtracting upstream irrigation flow from measured flow at 117 produced negative flows during parts of summer 2007, 2008, and 2009. These negative values were reset to zero to estimate natural flow magnitude (Figure 41). Subtracting the irrigation flows is less appropriate for this site (and site 103) than the upper canyon sites, because much of the irrigation was applied a considerable distance upstream from site 117, allowing processes like evapotranspiration and bank storage/shallow groundwater dynamics to potentially complicate the simple approach of determining natural flow rates by subtracting irrigation from the measured flows.

If irrigation water is included, ocular comparison of flows at reference site 185 and site 117 show a generally consistent pattern throughout the entire monitoring period, from 1994 through 2011; flow at 117 is always substantially greater than flow at 185 except for the summers of 2002 and 2003 when both sites experienced zero flow (Figure 40). This consistency of pattern suggests no construction impact at 117 from the City Creek/1999 mining. The pattern also suggests that with irrigation, flows at 117 were generally within the normal pattern, at least compared to reference site 185, through the entirety of the monitoring period. In other words, with irrigation flows at 117 were adequate at the coarse scale of an ocular comparison. However, a different pattern emerges when irrigation flows are subtracted out. During the summers of 2007 through 2010 flow at 117 hit zero when flow at 185 was always above 0 (Figure 41), implying a construction impact because this timing of reduced flows at 117 (compared to 185) is within the early 2006-to-early 2007 time range of impacts at three critical upstream wells—909, 955, and 956.

Baseflow comparison between sites 185 and 117 also suggests an impact at 117, particularly during 2007 and 2008, when 0 (or near 0) flows at 117 occur when flow at 185 was in the 1 to 2+ gpm range (Figure 42). This suggestion of an impact is not as clearcut as at some other sites because the assessment is 31

weakened by the absence of flows at 185 in the 1.5-to-2.3 gpm range. On balance I believe an indirect (up-channel driven) impact occurred at site 117. Rebound appears to have begun some time in summer 2010 and continues through September 2012.

Site 103—Located a few hundred feet north of the NFS boundary with the San Manuel Reservation, site 103 is the most distant Sand Canyon site from the tunnel alignment, approximately 0.9 mile from the northern segment of the alignment and 1.2 mile from the eastern/City Creek segment. Site 103 has a similar flow regime to site 117; perennial during the pre-construction duration of record except for brief no-flow periods in the summers of 2002 and 2003 (Figure 43). Any construction-related impact at site 103 is also probably indirect, caused by reduced flow from up-canyon. The considerations for the impact assessment at site 117 hold for site 103 as well: uncertainty in the natural flow rates because much of the applied irrigation water originated a distance upstream. However, a difference between the two sites is the sporadic irrigation—sometimes over 30 gpm—immediately above site 103.

If irrigation water is included, ocular comparison of flows at reference site 185 and site 103 show a generally consistent pattern throughout the entire monitoring period, from 1997 through 2011; flow at 103 is always substantially greater than flow at 185 except for the summers of 2002 and 2003 when both sites experienced zero flow (Figure 43). This consistency of pattern suggests no construction impact at 103 from the City Creek/1999 mining. The pattern also suggests that with irrigation, flows at 103 were generally within the normal pattern, at least compared to reference site 185, through the entirety of the monitoring period. In other words, with irrigation flows at 103 were adequate at the coarse scale of an ocular comparison. However, a different pattern emerges when irrigation flows are subtracted out. During the summers of 2007 through 2010 flow at 103 hit zero when flow at 185 was always above 0 (Figure 44), implying a construction impact because this timing of reduced flows at 103 (compared to 185) is within the early 2006-to-early 2007 time range of impacts at three critical upstream wells—909, 955, and 956.

Baseflow comparison between sites 185 and 103 also suggests an impact at 103, particularly during 2007 and 2008, when 0 (or near 0) natural flows at 103 occur when flow at 185 was in the 1 to 2+ gpm range (Figure 45). This suggestion of an impact is not as clearcut as at some other sites because the assessment is weakened by the absence of flows at 185 in the 1.5-to-2.3 gpm range. On balance I believe an indirect (up-channel driven) impact occurred at site 103. Rebound appears to have begun some time in summer 2010 and continues through 2012.

Little Sand Canyon

The Little Sand Canyon watershed is somewhat distinctive in being more ―long and narrow‖ compared to the other watersheds in the project area (Figure 1b). Like most of the other watersheds, however, Little Sand is a southwest-facing catchment occupying approximately 900 acres. The tunnel alignment crosses the watershed about one-third ―down‖ from the northern border, with most of the surface water and well monitoring locations south of the alignment. Flow in the main channel of Little Sand Creek becomes perennial at the upper main stem monitoring site, 44. Tributaries to the main channel do not flow perennially. Biological surveys in 2003 identified much of the lower main channel to be habitat for Least Bell’s Vireos and Southwest Willow Flycatcher and a nesting pair of Vireos was identified in the lower canyon on private land in 2005. Western Spadefoot toads were also seen during the surveys.

Five surface sites have been monitored by MWD staff during the ATP (Figure 1b). The northern-most, site 637 is in an intermittently-flowing section of the main stem, about 0.1 mile south of the alignment. Site 510, 0.4 mile south of the alignment, is also a stream site located on a tributary high above the main

channel. This site is of interest in its role as a water source for biota in the higher elevations of the watershed. Site 44 is a developed spring, actually an adit, where flow is measured a few feet below the seepage point from the bedrock. Site 509, on the main channel, is a few hundred below site 44. Main stem site 155 is further down the canyon at a relatively low gradient location where the canyon opens up. Distances from the alignment are 0.7 mile for site 44, 0.75 mile for site 509, and 0.95 mile for site 155. Given it’s relatively large distance from the alignment, any tunnel construction impacts to site 155 probably include an indirect, upstream component.

Four wells provide data relevant to ground and surface water dynamics in Little Sand Canyon. Water levels in well 954, located about 0.05 mile north of the alignment, were impacted by tunnel construction late in January 2006. Well 909, located on a west-facing hillslope close to the main channel, is about 0.37 mile south of the alignment. This well was impacted by tunnel construction initially late December 2005. Wells 958 and 178, respectively 0.95 and 1.05 mile south of the alignment, were not impacted by tunnel construction. Sites 637 and 510 are 0.1 to 0.15 mile from wells 954 and 909 respectively. Sites 44 and 509 are about 0.3 mile north of well 958 and about 0.35 mile south of well 909. Site 155 is equidistant—about 0.1 mile--from wells 958 and 178.

Site 637—Although the location of this site, about 0.1 mile from the tunnel alignment, and a similar distance from impacted well 954, would suggest a high potential for construction-related impact, the intermittency of flow at site 637 precludes confident impact assessment for this site. The vast majority of reported flows at 637, both before and after the early 2006 impact date for well 954, were zero, and do not allow assessment by the techniques used here (Figure 46). The impact status for site 637 is therefore indeterminate.

Site 510—Flow at this site was perennial from the beginning of the monitoring period in 1995 through mid-June 2001 (Figure 47). Thereafter no flow was measured until winter 2005. Much of summers 2006 and 2007 again had no measurable flow as did the period from November 2008 through the end of record in mid-2011. Irrigation water was applied at site 510 almost continuously from mid- February 2006 through mid-December 2008, usually at 2-3 gpm, and again from April through December 2009, at a much lower rate, 0.1 to 0.2 gpm.

Prior to summer 2001, flows at sites 510 and 185 were very similar (Figure 47). Through 2002-2004 flow at 510 was consistently zero while flow at 185, though often 0.5 to 2 gpm, was not zero during winter and spring. Flows at 510 again dropped to zero in February 2006 while flows at 185 were generally in the 2-3 gpm range. The reduced flow at 510 (relative to flow at 185) starting in February 2006 suggests a potential construction-related impact, particularly because nearby well 909 was impacted beginning in December 2005. Low summer baseflows at 510 compared to baseflows at reference site 185 (Figures 48 and particularly 49) suggest an impact at 510 every year from 2006 through 2010. On this basis site 510 is interpreted to be impacted with no evidence for rebound through May 2011, when monitoring at 510 ceased.

Site 44—This site, along with site 45 in Borea Canyon, differs from all other surface water monitoring locations in that these sites were developed historically as water supplies. The flow measurement is made immediately adjacent to the adit so that no contribution to the flow from surface runoff is possible. Similarly evapotranspiration effects on flow are non-existent; flow at site 44 is completely driven by subsurface/groundwater processes. Consequently there are no winter flow spikes common to stream monitoring sites. Flows at site 44 correlate very well with proximate well water levels and the flows at 44 vary much less than flows at most other sites (e.g., historical minimum = 2.5 gpm and historical maximum = 20 gpm). Sites 44 and 45 provide the best surface water proxies in the project area for groundwater dynamics. 33

Flows at site 44 can be categorized on a coarse scale into two phases: 1) relatively constant flow from early 1995 through late 1999, typically varying between 5 and 15 gpm and including an elevated period from mid-1998 through spring 1999 probably associated with elevated groundwater levels; 2) two periods of declining, recessional flow, first from spring 1999 through January 2005, and second from December 2005 through December 2010 (a third recessional flow period may have started in January 2011). Higher flows through most of 2005 and again through 2011 to date mimic the elevated flows of the late-1990s and are presumably due to elevated groundwater levels.

Visual comparison of flows at 44 and reference site 185 shows a narrowing of the differential (for flows between the two sites) from early 2006 onward compared to the differential prior to 2006 (Figure 50). This narrowing is not as large as in many other sites (e.g., site 56 in the City Creek watershed—Figures 3 and 4). Late 2005/early 2006 is, however the timeframe for the initial construction-related impact at well 909, approximately 0.35 miles from site 44. Note however, that the closest well to site 44, well 958 was not impacted (958 is about 0.3 mile further away from the alignment than site 44). Baseflow comparisons do show that the 2007-2010 flow relationship between sites 44 and 185 is outside the band of pre-2006 flows (Figure 51), indicating a potential construction impact at 44, particularly in 2009 and 2010.

Because even the 2009 and 2010 points on Figure 51 are not substantially removed from the pre-2006 baseline range of flow pairings (e.g., ~2 gpm below the lowest range of the pre-2006 relationship), comparison to a second reference site was attempted. Unfortunately, many sites on the Arrowhead east alignment are potentially impacted, reducing drastically the number of candidate reference sites. Summer baseflows at site 645, in upper Strawberry Creek, were compared to baseflows at site 44. Unfortunately, an insufficient number of pairings on the same measurement date (or within three days of measurement) were available to provide an adequate pre-2006 baseline so use of this second reference site isn’t workable. On balance, the evidence suggests a slight, but definite impact at site 44. There is good evidence of rebound in summers 2011 and 2012.

Site 509—This stream site is located a few hundred yards below site 45, on a low-gradient reach of the main Little Sand Canyon channel (Figure 1b). At about 0.72 miles south of the tunnel alignment, site 509 is also 0.4 mile south of well 909, and one-quarter mile north of well 958. Flow at 509 has been perennial throughout the complete period of record with the lowest flow (8 gpm) recorded in September 2009. Annual minimum flows have otherwise been in the 20-30 gpm range since 2003, and 35-50 gpm earlier.

Visual comparison of flows for site 509 and reference site 185 does not show an obvious change (narrowing or widening) at any particular time (Figure 52), suggesting no evidence of a construction impact. Baseflow comparison, however, shows some 2008, and most 2009 and 2010 flows to be outside the range of most flows occurring prior to the late 2005/early 2006 impacts at wells 954 and 909 (Figure 53) (One pre-2006 flow pairing, at 3, 34 extends the range of the pre-2006 point ―envelope‖ downward. The 34 gpm flow at site 509 on September 23, 1999 was the lowest flow on record prior to the dry period starting in 2002. This flow measurement is not necessarily an anomaly but it is somewhat of an outlier). A construction-related impact at site 509 is therefore identified for part of 2008, and 2009 and 2010. Rebound was underway in summers 2011 and 2012.

Site 155—Monitoring site 155 is near the base of Little Sand Canyon, on private land a few hundred feet upstream of a location where nesting Least Bells Vireos were observed during the construction phase of the project. This site is almost one mile south of the alignment, about 0.12 mile west of well 958 and about 0.1 mile north of well 178. Flow at 155 has been perennial through the entire

period of record, with annual minimums often above 35 gpm through 2001, 20-30 gpm from 2002 through 2004, and below 10 gpm in 2009 and 2010.

As with site 509, visual comparison of flows for site 155 and reference site 185 does not show an obvious change (narrowing or widening) at any particular time (Figure 54), suggesting no evidence of a construction impact. Summer baseflow comparison, however, shows 2008 through 2010 flows in particular to be outside the range of flows occurring prior to the late 2005/early 2006 impacts at wells 954 and 909 (Figure 55). Baseflows at 155 in 2006 and 2007 are also generally outside the range of pre-2006 baseflows. Because of site 155’s distance from the alignment, and it’s proximity to two wells that were not impacted by tunnel construction, the impact at 155 is interpreted to be indirect, and a result of reduced/impacted flows upstream that translate down to site 155. In 2011 and 2012 there is evidence for some rebound in flows, although the flow relationship for both years is not completely within the pre- 2006 envelope.

Borea Canyon

Occupying about 700 acres, Borea Canyon is the smallest of the three main canyons on the eastern alignment of the project area (Figure 1b). The tunnel alignment bisects the canyon into northern and southern halves. Several tributaries enter the main canyon but flow at all of them is ephemeral. Flow at both surface water monitoring sites in Borea Canyon was perennial during the period of record prior to tunnel construction. Site 45, an adit (―horizontal well‖ in MWD’s parlance) that is effectively a true spring, is located about 0.4 mile south of the alignment. Flow in the main channel above site 45 is ephemeral, with little riparian vegetation evident in the main channel immediately above site 45. Flow at 45 dropped precipitously in summer 2005, to 0.01 gpm before rising that autumn. In October 2008 flow dropped to 0 and did not rise beyond 1 gpm until February 2010. At site 154, about 0.55 mile south of the alignment, flow continued perennially during the complete tunnel construction period. This monitoring site is located in the main channel of Borea canyon a few yards below a seep. 154 therefore receives water from both a spring (seep) and from upstream via the main channel. Biological surveys in 2003 identified much of the lower main channel to be habitat for Least Bell’s Vireos and Southwest Willow Flycatcher.

Three wells provide water level data relevant to potential tunnel construction effects in Borea Canyon. Well 907 is on the ridge forming the western boundary of the Borea watershed. It’s about 0.12 mile south of the alignment, one-half mile north northwest of surface site 154, and about 0.35 mile northwest of site 45. A construction impact at well 907 first appeared in February 2004. Well 908 is on a hillslope west of the upper main channel, about one-quarter mile south of the alignment, 0.2 mile north of site 45 and 0.35 mile north of site 154. Water levels in well 908 were impacted starting in mid 2004. Well 953 is the closest of the three Borea wells to the alignment, about 0.10 mile north. With its northerly position, 953 is the most distant from the surface water sites—about one-half mile north of 45 and two-thirds mile north of 154. Well 953 was first impacted in mid-2004.

Site 45–Flows at site 45 are similar those at Little Sand Canyon site 44 in being completely controlled by subsurface geo-hydrological processes (the monitoring point is isolated from any source of surface runoff--although potentially surface runoff could contribute to flows at 45 by percolating into the ground and moving in shallow sub-surface pathways to the adit). As with 44, the hydrograph of 45 mimics that of a nearby well, 908, with almost no evidence of mid-winter rain-induced spikes (Figure 56).

Prior to summer 2004 flow at 45 was always above 10 gpm. Thereafter the 2001-to-2004 period saw a steady decline culminating in early December 2004 with flow dropping to 2 gpm (by far the lowest on

record and without a concurrent extreme decrease in flow at reference site 185). Mitigation water was about to be applied in early December 2004 when unusually high and intense rainfall elevated flow at both sites. The flow increase was relatively short-lived at 45 and flow declined quickly from early March to zero in summer 2005. 2006 flows at 45 rebounded somewhat but did not regain the pre-2004 differential versus site 185 flow. 2008 and 2009 flows at 45 were the lowest on record, with zero flow measured routinely from late 2008 through late 2009. Rebound started in 2010 and continuing through the beginning of summer 2011.

Baseflow comparisons further illustrate the change in flow relationship between sites 185 and 45 (Figure 57). The pre-2004 ―baseline‖ relationship (brown squares) is not maintained during the summers of 2005- 2010, inferring a shift in the relationship. Construction impacts at the three Borea wells ranging from early to mid-2004 coincide with the relatively severe drop in flows at surface site 45. Taken together the timing of the well water declines, the baseflow, and through-the-year flow comparisons between sites 185 and 45 suggest a high likelihood of a construction effect occurring at site 45 in summer 2004.

Mitigation water was continuously applied at a location 30 yards above 45 at approximately 1-2 gpm from 7/8/05 through September 2006, then upped to 2-3 gpm through August 2008. Because of vegetation dieback further down canyon mitigation water was applied between sites 45 and 154 at 1 gpm at the new mitigation site and about 4 gpm above 45 from 9/9/08 through 12/12/08 when application of mitigation water ceased at both locations. Irrigation was resumed 4/29/09 solely above site 45 and ranged from 3 to 5 gpm until suspension on 12/7/09. Although natural flow resumed at site 45 in summer 2010, irrigation at 1.5 to 2 gpm was applied from late July 2010 through late October 2010 (none of the 2006-2011 ―points‖ on Figure 55 include irrigation water). Flow at 45 in summer 2011 was consistently 21 gpm, indicating a rebound, although neither the summer 2011 nor summer 2012 flows are completely within the baseline (pre-2004) relationship.

Site 154–Flow at site 154 exhibits characteristics of both a spring and a stream, with more variable flow than at site 45 stemming from rain/surface runoff-induced winter elevated flows, superimposed on more steady spring flow (Figure 58) (note that all plotted flows at site 154 include any residual irrigation water applied higher up the canyon during much of the period from 2005 through summer 2010). As with several other sites in the project area, flows at 154 dropped over the 2001-2004 dry period and at 154 reached near-zero historical lows in December 2004. Highest-on-record flows in winter 2004-2005 were then followed by flows through summer 2010 that very seldom reached pre-2004 levels. During the late-2004 through 2010 period the flow differential between sites 185 and 154 narrowed relative to the pre-2003 differential (Figure 58). The December 2004 minimum flow at 154 suggests a potential construction impact beginning some time in summer or autumn of 2004.

Baseflow comparison between site 154 and reference site 185 confirm a distinct shift in the flow relationship before 2004 and from 2004 through 2010 (Figure 59). The plotted flows in the figure for site 154 are potentially augmented by irrigation water from the upper canyon. Therefore if anything the 154 flows in Figure 59 are high during the summers with irrigation, further emphasizing the disparate nature of the baseflow relationship before 2004 and after. As with site 45, well impacts in 2004 coincide with the change in surface baseflow relationship, suggesting a high likelihood of a construction effect at site 154 occurring some time in 2004. Rebound at 154 began in 2011 but was not complete as of September 2012.

Strawberry, East Twin, and Waterman Canyons

The surface water monitoring sites in Strawberry, East Twin, and Waterman Canyons are grouped together here because these watersheds share characteristics that differ from the other ATP area watersheds. Specifically these catchments drain much larger watersheds including land ranging up to the top of the San Bermardino Mountain plateau, and consequently incorporate hydrological attributes that differ decidedly from the lower elevation ATP watersheds. The larger nature of these watersheds mean that any potential impacts from tunnel construction are probably muted, simply because of the larger, and more distant from the alignments, land area that contributes to hydrological processes. The combined East Twin/Strawberry drainages occupy over 5,630 acres. Waterman Canyon watershed is smaller, occupying approximately 3,200 acres.

Twenty surface water monitoring sites were operational in the three watersheds: five (191, 93, 642, 94 and 134) near the eastern portal of the western tunnel; another four (624, 189, 38 and 201) on lower East Twin Creek; one (90) at a hot spring about 700’ west of Arrowhead Springs hotel, two (628 and 644) on middle and upper E Twin Creek respectively; 629 on lower Strawberry Creek; a cluster of five (120, 676, 677, 678, 679) on middle Strawberry Creek, and the final site, 645, on upper Strawberry Creek (Figure 1a). The vast majority of these sites are north of the two alignments.

Six of the eight wells in these watersheds are located close to the eastern portal of the western tunnel. None are in the East Twin or Strawberry drainages. Wells 197, 905, 923, 946, and 918 are all within a couple hundred feet of the tunnel alignment, with well 937 located about 1400’ southwest of the portal (Figure 1a). Construction-related impacts at wells 197 and 923 began in late April 2004. Monitoring at well 905 went dry in November 2003, almost concurrent with the beginning of mining. There’s no evidence of construction-related impact at wells 946, 918, or 937. The other two wells, 906 and 198, both located in Harris Canyon very near the eastern tunnel alignment, are dry, with monitoring at 906 ending in December 2003.

Site 191—Located in the lower portion of Waterman Canyon, site 191 monitors the main channel, approximately 0.15 mile north of the western tunnel alignment (Figure 1a). Flow at 191 has been perennial throughout the entirety of the period of record except for the summer of 2002, when three consecutive monthly measurements were all zero. Except for the dry years of 2000-2002 summer flows at 191 were typically above 100 gpm.

The relationship between flows at reference site 185 and site 191 does not appear to change through the entire period of record (Figure 60), suggesting no construction-related impact. The close proximity of site 191 and well 197, about 0.3 mile, would suggest that any construction-related impact at 191 approximate the time of impact to well 197, late April 2004. The flow pattern in Figure 60 does not change after April 2004 nor does a comparison of summer baseflows before 2004 versus the following summers suggest an impact at site 191 (Figure 61). Consequently site 191 is not deemed to have been impacted by tunnel construction.

Site 93 is located about 1300’ northeast of the western tunnel alignment on a tributary to lower Waterman Canyon (Figure 1a). This site monitors geothermal waters often with temperatures of 120F. Since the beginning of routine monitoring at this site in May 1995, non-zero summer flow occurred only in 1995 and 1998, before any possibility of a construction-related impact. Consequently data during the most critical part of each year, the summer baseflow period, isn’t available during the period of potential impact. Therefore the reference/impact technique used here isn’t applicable to this site, and the impact status is indeterminate. Nevertheless, even though this site is close to the alignment, the frequent zero flows both before and after a potential spring 2004 impact date suggest a low likelihood of impact.

Site 94 is also located close to the alignment (about 1200’ distant), and is also about 1200’ from impacted well 923 (Figure 1a). Like site 93, flow at this location is thermally influenced, with temperatures in the 180-190F range for almost every measurement. This site is atypical in that flow emanates from a horizontal pipe. Flow magnitude at 94 is atypical in that the flow variation has been minimal since summer 1999 (Figure 62). Because flow at reference site 185 changes through time there’s no ocular evidence of a consistent relationship between flows at 93 and 185 during any time period, and therefore no evidence for a changed relationship before versus after the April 2004 time of potential construction-related impact (Figure 62). Comparison of summer baseflows alone between the two sites also shows no evidence of a changed relationship pre- versus post-2004 (Figure 63). On these bases site 94 was not impacted by tunnel construction.

Site 642 is located on private land (Campus Crusade) about 900’ east of the eastern portal of the western tunnel (Figure 1a). The location of this site suggests that its potential impact date would approximate the beginning of mining on the western tunnel, October 2003 at the earliest.

Prior to February 2004 flows at this site were perennial and never lower than 0.5 gpm during the period of record (Figures 64 and 65). And flow was relatively constant throughout each year, with little spiking in winter and spring, suggesting that this site is primarily groundwater-dependent. In February and March 2004 zero flow was measured. This zero flow is atypical for February and March of any year at most sites. There is a definite ocular shift in flows between sites 642 and 185 starting in early 2004. Prior to 2004 winter flows at 185 generally approximated or were higher than those at 642 (Figure 65). During the 2004 through mid-2006 period flows at 642 were consistently lower throughout each 12-month period than at site 185, a definite departure from the pre-2004 pattern. Unfortunately few baseflow measurements at 185 and 642 were made within a three-day window. Consequently baseflow comparison is not possible. Comparison of flows at 642 against a second reference site, 645, also ocularly suggests a change in pattern, in 2004 when flow at 642 dropped to zero without a similar drop at 645 (Figure 66). Again, unfortunately there are insufficient concurrent (or near-concurrent) summer baseflow measurements at 642 and 645 to allow baseflow comparison. The shift in flow patterns soon after initiation of tunnel construction suggests a construction-related impact at this site, but without baseflow comparisons this conclusion cannot be substantiated, and on balance a ―gray area‖ determination is made for site 642.

Site 134—Located within a few hundred feet of the eastern portal of the western ATP tunnel, site 134 would be expected to be impacted by tunnel construction no earlier than the October 2003 date of initial mining (Figure 1a).

At site 134 flows were have been perennial throughout the entirety of the period of record, with summer lows typically in the 2-4 gpm range (Figure 67). Ocular comparison of flows between site 134 and reference site 185 do not show an obvious change in pattern; flows at 185 are almost always proportionally lower than those at 134 throughout the full period of record. Baseflow comparisons also do not identify any change in pattern between sites 134 and 185 throughout the entirety of the data record; the pre-2004 range of flows is within the 2004+ range of flows (Figure 68). Consequently site 134 was not impacted by tunnel construction.

Site 90—Although MWD records flow from this site, water actually emanates from a concrete- capped well, not a spring, stream or otherwise a natural landscape feature. Also, flow at this site has been monitored at a 6-month frequency for the entire span of record. This monitoring frequency is too seldom to provide any basis for a supportable impact analysis. Consequently there’s no way effectively to determine the impact status at this site.

Site 644—At over one mile north of the west tunnel alignment, this site is the most distant within the East Twin drainage (Figure 1a). The site monitors the main channel of E. Twin Creek where flow is perennial, with summer minimums usually at least 100 gpm (Figures 69 and 70). Well 907 is the nearest well to site 644; it was impacted in February 2004. There’s no ocular change in pattern between flows at site 644 and reference site 645 during any portion of the period of record (Figures 69 and 70); the proportional difference in flow between the two sites does not vary appreciably through time. Baseflow comparison for flows prior to 2004 and annually from 2004 through 2010 also does not identify a changing relationship; the 2004-2010 flows are within the envelope of pre-2004 flows (Figure 71). However, the baseflow comparison at site 644 is atypical in that flows after the baseline period (e.g., in 2005 and 2006) were greater than any flows during the baseline period (e.g., pre-2004). This means that the baseline ―envelope‖ of flows doesn’t contain the 2005 and 2006 flows; extrapolation to the higher flows is risky and the relevancy of the 2005 and 2006 flows to the pre-2004 baseline period is uncertain. Nevertheless, even given the uncertainty of the status of the 2005 and 2006 flows, they did not drop precipitously or otherwise clearly suggest a changed flow relationship. Consequently there’s no evidence that site 644 was impacted by tunnel construction.

Site 628—Also on the main stem of East Twin Creek, site 628 is about one-half mile north of the eastern tunnel, and approximately equidistant from Borea Canyon well 907 and the eastern (Waterman) portal of the western tunnel (Figure 1b). During the period of record, from 1999 through 2010, flow was generally perennial at this site. However, measurements in August of 2000 and August 2001 recorded zero flow (Figure 72). Site 628 could potentially be affected by tunnel construction activities in October 2003 (the beginning of mining for the eastern tunnel) and/or February 2004, when well 907 was first impacted.

The comparative flow pattern for site 628 and reference site 645 does not change markedly over time; summer flows are generally lower at 628 with winter flows approximately equal at the two locations (Figure 72). On this basis there’s no evidence for a construction effect at site 628. Baseflow comparison also does not support a construction impact at this site; flows during the potential impact period (2004 and later) are within the envelope of flows prior to 2004 (Figures 73 and 74). However, the baseflow assessment is limited as with site 644—flows in 2005 and 2006 in particular were much higher than any during the pre-2004 period, adding uncertainty to the analysis for these years. On balance, there’s no clear evidence for a construction-related impact at site 628.

Site 201—According to mapping by MWD this is a spring site adjacent to the main channel of East Twin Creek, below it’s confluence with Strawberry Creek (Figure 1b). Flow at 201 dropped to zero during the dry summers of 2000 through 2003 and otherwise was generally in the 2-15 gpm range prior to early 2004. Per MWD employee Tom Hibner, mass movement (―land sliding‖) associated with the December 2003 flooding in the ATP area covered the monitoring location and consequently changed the flow regime at site 201. All measurements after December 2003 have been zero. This site is about 0.3 mile north of the eastern tunnel and about one-half mile from the portal of the eastern tunnel. From this location any potential construction-related impact to flows at site 201 would be expected between the initiation of mining for the eastern tunnel—mid-August 2003—and the February 2004 initial impact at well 907. Any potential construction-related impact at site 201 would have begun no earlier than approximately when the site was covered by rock debris in December 2003. Consequently it’s impossible to assess the impact at this site because the flow regime changed and the baseline information would not be relevant to the post-baseline/candidate impact period.

Site 38 is adjacent to site 201 on the main channel of East Twin Creek immediately south of the Arrowhead Springs hotel complex (Figure 1b). Flow at 38 is perennial with summer minimums typically in the 10-100 gpm range (Figure 75). This site is about 0.3 mile north of the eastern tunnel and about 39

one-half mile from the portal of the eastern tunnel. From this location any potential construction-related impact to flows at site 201 would be expected between the initiation of mining for the eastern tunnel— mid-August 2003—and the February 2004 initial impact at well 907.

The relationship between flows at site 38 and reference site 185 does not change obviously over time, including before versus after the late 2003/early 2004 time frame when a construction-related impact might be expected to begin (Figure 75). The summer baseline baseflow relationship between these two sites is not ―tight‖, particularly at the low end of the flow scale where flow at 38 does not vary appreciably with flow at 185 (through 4+ gpm) (Figures 76 and 77). This produces a wide band for the baseline flows and consequently lowers the reliability in the use of the reference-impact technique for this site. Unfortunately no other reference site is available with a tighter baseline relationship. Given this caveat, there’s no substantial evidence for a construction-related impact at site 38, although again the assessment for this site is relatively weak.

Site 189 is very close—about 0.05 mile—to the western portal (Strawberry) of the eastern tunnel. Flow on this section of E Twin Creek is perennial, with summer minimums generally above 200 gpm (Figure 78). Given its close proximity to the tunnel portal, any construction-related effect would be expected to begin no earlier than concurrent with initial tunnel construction in October 2003.

The relationship between flows at site 189 and reference site 645 does not obviously change through time; flow at 645 is proportionally less than flow at 189 throughout the entirety of the 1999-2010 period of concurrent measurements (Figure 78). Baseflow comparison, however, shows summer flows at site 189 in parts of 2007 through 2009 to be outside the range of the pre-2004 flows (Figures 79 and 80). Flows in 2004 are not outside the pre-2004/baseline flow range (the one 2005 flow pairing is of a much greater magnitude than the baseline flows, and consequently less reliable an indicator of impact--or lack of impact). There’s no clear reason why a construction impact would become apparent in 2007, several years after the logical time frame—2004—for any anticipated impact. On this basis the impact status for site 189 is ―gray‖; there’s evidence for a changed flow relationship but at a time period that does not make sense for a construction-related impact.

Site 624 monitors flow in the main channel of lower E. Twin Creek, very close to the bottom of the mountain front (Figure 1b). Flow at 624 has been perennial throughout the period of record with minimums in the 100-300 gpm range during dry summers (Figure 81—note division by 1000 on graph for flows at site 624). Site 624 is about 2000’ south of the western portal (Strawberry) of the eastern tunnel. Given its proximity to the tunnel portal, any construction-related effect would be expected to begin no earlier than concurrent with initial tunnel construction in October 2003.

The relationship between flows at site 624 and reference site 185 appears to change through time, from mid-2005 through early 2008 flows at site 624 are proportionally lower compared to flows at site 185 than prior to mid-2005, suggesting a potential construction-related impact beginning possibly in mid- 2005. Comparison of summer baseflows prior to 2004 and afterwards does not, however, substantiate a changed relation; baseflows from 2005 through 2010 are within or above the range of pre-2004 flows (no measurements were made within 3 days at the two sites in 2004 or 2011). On this basis there’s no evidence for a construction-related impact at site 624.

Site 645—At 1.5 miles north of the alignment of the eastern tunnel on the main stem of upper Strawberry Creek, site 645 is the most distant of any of the 70 plus surface water monitoring sites from either tunnel (Figure 1b). This distance alone reduces the likelihood of a tunnel impact at this location. Flow at site 645 has been perennial throughout the duration of the 1999 through 2010 period of record

(Figures 69 and 70), although summer minimums have been below 5 gpm. The closest well to site 645 is 954, which was initially impacted in late January 2006.

As mentioned above in the Methodology section of this report, identifying reference sites in the ATP area is not easy, particularly on the eastern alignment where all parties concerned agree that flows many locations have been impacted by tunnel construction activities . Given the desirability of baseflow measurements to be within a few days at reference and non-reference sites, and baseflows that are approximately equal between reference and non-reference sites, 645 is used as a reference for many of the Waterman/E Twin/Strawberry monitoring sites. Designation as a reference is based on 645’s distance from the tunnel, its approximately equivalent flow magnitude to many Waterman/E Twin/Strawberry monitoring sites and to the lack of any obvious signs of impact.

To further explore the potential for construction-related impact at site 645, flows there are compared to flow at the USGS stream gage at the base of Plunge Creek. This pairing of flow records is not ideal. The Plunge Creek gage is over 7.5 miles southeast of site 945, and differing measurement techniques are used at the two sites, potentially adding variability to the measured flow amounts. Flows at the two locations are generally proportional throughout the duration of record (Figure 83), with flows at 645 almost always considerably lower than at Plunge Creek. Results from summer baseflow comparison between these two sites are somewhat ambiguous. Most of the available baseflows from the two sites during the pre-2006 (baseline) period are lower than flows from 2006 through 2010 (Figure 84). This leaves only five baseflow pairings within the range of magnitude of most of the 2006 through 2010 flows, a small number to base conclusive results on. Some flows in 2008 through 2010 are outside the range of the baseline period in Figure 84, suggesting a changed flow relationship. There’s no obvious reason, however, to expect tunnel construction as a possible cause for this possible changed relationship; tunnel impacts—as in the vast majority of sites that are impacted—are expected to occur within a few months of the anticipated date, in this case January 2006. On the basis of distance from the tunnel and the lack of a confirmed ―link‖ between a changed baseflow and tunnel construction no tunnel-related impact is identified for site 645.

Sites 120, 676, 677, 678 and 679—These sites are grouped because of their close proximity in upper Strawberry Creek and their shared flow and data characteristics. All five sites are about 4000’ north of the eastern tunnel alignment (Figure 1b) and about the same distance from the nearest well, 953. Data collection at all of these sites began in late 1999, providing four years of potential baseline data prior to the mid-2004 construction impact at well 953. The 2000-2003 sequence of years was particularly dry, and most of these sites had zero summer flow, leaving few non-zero flows on which to construct baseline relationships. Also many 2004-2010 summer flows were zero (e.g., at site 677 zero flow for all summer measurements in 2004, 2007 through 2010, and two of four 2006 measurements). Last, because of the low (when non-zero) flows during the baseline period, any baseline relationships that could potentially be developed did not incorporate high enough flows relevant to summer flows in 2005 and 2006; in other words a risky extrapolation would be needed to address flows during the more normal summers of 2005 and 2006. For these reasons it’s not possible to use the reference-impact technique to determine impact status at sites 120, 676, 677, 678 and 679.

Site 629—On the lower section of the main channel of Strawberry Creek, site 629 is about 0.55 mile north of the eastern tunnel alignment (Figure 1b). Flow has been perennial at 629 throughout the entirety of the available data record, with summer minimums typically in the 20-60 gpm range (Figure 85). Well 907, the closest to site 629, was initially impacted by tunnel construction in February 2004. February 2004 was also the date of the tunnel boring machine’s closest approach to site 629.

Comparison of flows through time at site 629 and reference site 645 does not identify any change in the flow relationship (Figure 85); flows are similar between the two sites. Summer baseflow comparison between the two sites also generally shows flow pairings for 2005 through 2010 to be within the range of pre-2004 flows (Figures 86 and 87). Two flow pairings, one each for 2008 and 2009 are outside the pre- 2004 range. As with other sites in this part of the ATP area, it’s not clear that a construction-related impact would occur in 2008 or 2009, over four years after passage of the tunnel boring machine. Consequently there’s no definitive evidence for a construction-related impact at site 629.

Unnamed Face Watersheds Located Between Waterman and Sycamore Canyons

Three surface water monitoring sites, 65, 17 and 627, are located in small ―face‖ watersheds between Waterman and Sycamore Canyons. These watersheds drain directly onto the alluvial piedmont at the base of the San Bernardino Mountains. The closest well to these surface sites, well 903, located several hundred feet east of the watershed divide with Sycamore Canyon and very near the tunnel alignment, was initially impacted in mid-July 2005.

Site 65, at 2190 ft altitude, is a spring located on private property at the ―U turn‖ on highway 18, about 0.4 mile south of the alignment (Figure 1b). This site flowed perennially in the 5 to 25 gpm range from 1994 through mid-2001 when flows dropped typically to the 0.5 to 7 gpm range through summer 2005. At that time flows dropped again to typically less than 1 gpm and to 0 in July 2006. Site 65 was irrigated, usually at 1-1.5 gpm, from mid-September 2006 through mid-January 2007, and again from June through mid-February 2008.

Site 17 is located at 2085 ft altitude, a few yards from highway 18 near the ―U turn‖ and a mile from the Arrowhead Springs Hotel. The small watershed draining into site 17 is bisected by the upper arm of the ―U‖ (Figure 1b). Flow at 17 was zero during parts of summers 2002 through the end of monitoring in late January 2007. Monitoring was suspended early in 2007 at the insistence of the private property owner. Irrigation was underway at site 17 from late July 2005 through late January 2007. The efficacy of the irrigation is incompletely known because each application of irrigation water was an instantaneous ―dump‖ from a truck rather than a continuous input. The author observed irrigation water sluicing down the channel over a period of no more than one-half hour; without the irrigation water the channel was dry. Site 17 is also about 0.4 mile south of the alignment.

Site 627, at 1590 ft altitude, is on an unnamed intermittent channel at the base of mountain front (Figure 1b). This site is about 0.8 mile south of the western alignment. Flow at 627 is intermittent, with zero flow during each summer during the entire monitoring period.

Site 65—Comparison of flows through time at site 65 and Ben Canyon reference site 11 identify a changed flow relationship starting in February 2005 when flows at site 65 dropped drastically compared to flows at site 11 and remained near zero for much of the time until February 2010 (Figures 88 and 89). An early 2005 date for construction-related impact to site 65 is earlier than the mid-July 2005 date of initial impact at well 903, possibly because site 65 is located east of well 903 and would theoretically therefore be impacted sooner.

Baseflow comparison between site 65 and reference site 11 confirm a distinct shift in the flow relationship before 2005 versus from 2005 through 2008 (Figure 90). The decrease in flows at site 65, roughly coincident in time with an impact at nearby well 903, suggest a high likelihood of a construction effect at site 65 occurring during the first few months of 2005. Rebound at 65 appears may have

occurred by 2010 (Figure 89), but with only one baseflow data point available in 2010 and none in 2011 (Figure 90)—because of a lack of concurrent measurement at sites 11 and 65--this conclusion is somewhat tentative.

Site 17— Comparison of flows through time at site 17 and Ben Canyon reference site 11 identify a changed flow relationship starting in early 2005 when flows at site 17 dropped compared to flows at site 11 and remained near zero for much of the time until the last measurement in January 2007 (Figures 91 and 92). An early 2005 date for construction-related impact to site 65 is earlier than the mid-July 2005 date of initial impact at well 903, possibly because site 17 is located east of well 903 and would theoretically therefore be impacted sooner.

Baseflow comparison between site 17 and reference site 11 confirm a distinct shift in the flow relationship before 2005 versus from 2005 through 2006 (Figure 93). The decrease in flows at site 17, roughly coincident in time with an impact at nearby well 903, suggest a high likelihood of a construction effect at site 17 occurring during the first few months of 2005. There’s no evidence in rebound in flows at site 17 before measurements were suspended in early 2007.

Site 627— Since the beginning of monitoring at this site in January 1998, non-zero summer flow occurred only in 1998, before any possibility of a construction-related impact. Consequently data during the most critical part of each year, the summer baseflow period, isn’t available during the period of potential impact. Therefore the baseflow comparison technique used here isn’t applicable to this site, and the impact status is indeterminate. Nevertheless, the frequent zero flows both before and after a potential early 2005 impact date suggest a low likelihood of impact. Also comparison of flows at site 627 and reference site 185 do not show a changed relationship; flows at 185 are always higher than those at 627, except for the dry period of summers 2002 through 2004 (Figures 94 and 95).

Sycamore Canyon

At 662 acres area Sycamore Canyon is the largest of the three drainages totally within the western tunnel alignment. Elevations range from about 1500’ at the canyon’s outlet east of the campus of San Bernardino State University, to 4003’ at Marshall Peak. The tunnel alignment bisects the watershed, with all of the surface water monitoring sites south of the alignment. Flow in the main channel of Sycamore Creek becomes perennial at the upper main stem monitoring site, 156. Tributaries to the main channel do not flow perennially. Stretches of the main channel were mapped in 2003 as Least Bell’s Vireo and Southwest Willow Flycatcher habitat.

Six surface sites have been monitored by MWD staff during the ATP (Figure 1b). All are on the main channel, and although all six experience brief winter flow spikes, the ―spikiness‖ is not as drastic as at some sites on the western alignment. The northern-most site, 156, at 2110 ft altitude and about 0.15 mile south of the alignment, flowed perennially prior to tunnel construction. 156 was irrigated, typically at 1 to 1.5 gpm, almost continuously from late October 2006 through late November 2008. Without the irrigation, it appears that there would have been zero flow in winter 2006-2007 and summer 2007. Surface flows at site 205, 0.23 mile south of the alignment at 2060 ft altitude, are perennial—with summer baseflows typically no less than 7 gpm--throughout the entirety of the monitoring record from late 1993 thru late 2011. At site 30, 2010 ft altitude and 0.28 mile south of the alignment, flows during the monitoring period were generally above 8 gpm. Site 20, on private land just south of the Forest border and 0.32 mile south of the alignment, flows in the 8-20 gpm range during summer, with winter spikes sometimes above 100 gpm. Springsnails were found at site 20. The lowest two surface sites in Sycamore Canyon, 182 and 95, are not assessed for potential construction-related impact because flow

measurements were influenced by water diversion near these sites starting in May 2004, long before any potential impact. Because the magnitude and timing of the diversions is unknown, the flow record at these sites is compromised in terms of impact assessment.

Water level data from three wells are potentially relevant to impact assessment for the Sycamore Canyon surface water sites (Figure 1b). The lower interval of well 952, located in the upper central portion of the watershed about 0.18 mile north of alignment, was definitely impacted by March 2006 and may have been impacted in July 2005, when well 903 was impacted. 903 is almost on the alignment a few hundred feet east of the Sycamore Canyon watershed divide. Well 902, like well 903, is located within a few dozen feet of the alignment, but just over the watershed divide into Badger Canyon. Well 902 was initially impacted in January 2007. All four of the surface water sites assessed here for tunnel-related impact are each approximately equi-distant from the three wells. However, given the east-west spread of the three wells, well 903 (the eastmost) would generally be expected to experience a tunnel-related impact before the surface sites, and well 902 (the westmost) after the surface sites. Well 952 is located almost directly north of the surface monitoring sites and as such its impact date would most likely approximate the date of any tunnel-related impact at the surface sites (Figure 1b), based on a simple as-the-crow-file distance between 952 and the surface sites.

Site 156--Visual comparison of flows at 156 and reference site 11 shows a narrowing of the differential (for flows between the two sites) from mid 2006 through late 2007 (Figures 96 and 97 in particular). This narrowing is not as large as in some other sites (e.g., site 56 in the City Creek watershed—Figures 3 and 4) but because irrigation water was applied during much of this period and included in the site 156 flow measurements, the differential is potentially larger than it appears visually. March 2006 is, however the timeframe for a construction-related impact at well 952, approximately 0.28 miles from site 156. Baseflow comparisons do show that the 2007-2010 flow relationship between sites 156 and 11 is outside the band of pre-2006 flows (Figure 98), indicating a potential construction impact at 156, particularly in 2006 and 2007. Because the evidence for a construction-related impact at site 156 is more tenuous than at some other sites, baseflow comparison was done with a second reference site, 27 in Badger Canyon. Results from this second baseflow comparison (Figure 99) provide stronger evidence for a tunnel-related impact at site 156, again particularly considering that flows at 156 include irrigation water in 2007 and 2008. There is some evidence for recovery at site 156 in 2010, although the single data point available is inadequate to be totally conclusive (Figure 98).

Site 205--The relationship between flows at reference site 11 and site 205 does not appear to change through the entire period of record (Figures 100 and 101), suggesting no construction-related impact. Flow at 205 is always greater than flow at 11 by approximately the same relative amount. As with the other Sycamore Canyon sites, any tunnel-related impact would be expected about when well 952 was impacted, initially no later than March 2006, or July 2005 at the very earliest. The baseflow relationship between sites 205 and 11 does not change from the baseline, pre-2005 period through the end of monitoring in 2010; data points for each year from 2005 through 2010 are within the range of the pre- 2005 points (Figure 102). Consequently site 205 is not deemed to have been impacted by tunnel construction.

Sites 30 and 20—The impact assessments for sites 30 and 20 are similar to the assessment for site 205; there’s no evidence for a tunnel-related surface flow impact at either of these sites on the basis of either a visual comparison of flows through the entirety of the monitoring record or a changed relationship in summer baseflows for years before and after a potential impact in 2005 (Figures 103 to 108).

Badger Canyon

Badger Canyon watershed differs from the other watersheds in the project area in having two channels through most of its length, with the channels joining near the base of the watershed (Figure 1b). The tunnel alignment bisects this 519-acre watershed. Watershed elevation ranges from 1740 ft at the base to 4000 ft at Marshall Peak. Western spadefoot toads and an adult Monterey Ensatina were observed in 2003 in the lower portion of the watershed, and habitat for Least Bell’s Vireo and Southwestern Willow Flycatcher was also identified in the lower half of the watershed at that time. Springsnails were monitored at two surface water monitoring sites in the upper portion of the watershed for several years.

Seven surface water locations have been monitored in Badger Canyon--sites 27, 213 and 214 located within a few yards of each other in the upper canyon north of the tunnel alignment, and sites 28, 26, 21 and 152 south of the alignment. The 27/213/214 cluster is about one-sixth mile north of the alignment with water at site 27 being the combined flow of water monitored immediately upstream in two tributaries. All three of these sites are at about 2800 ft elevation. Flow at 27 was perennial during the monitoring period except for zero flow during the summers of 2002 and 2003. Flows were also zero at site 214 during each summer from 2000 through 2003, but only during summer 2003 at site 213. Flow at all three of these sites was typically low, above the 5-15 gpm range only during winter spikes. Stream site 28, about one-tenth mile south of the alignment at 2375 ft elevation, flows ephemerally, with only nine non-zero measurements out of 250 made between December 1993 and March 2009. Site 26, located at 2280 ft on a small tributary east of the main Badger Canyon channel, is about one-fifth mile south of the alignment. Flow at 26 was perennial through June 2001 after which all measurements through the end of record in January 2003 were zero. Stream site 21, at 2100 ft elevation and the lowest site in the main canyon, is about 0.55 mile south of the alignment. Flow was perennial at this site until the last reported measurement in August 2001. At site 152, located at the mouth of Badger Canyon at the base of the mountain front, flow has been perennial, but lower since mid-2001. Measurements here were made every six months through mid 2006, typically each November and May. Thereafter measurements were typically weekly.

Four wells are proximate to Badger Canyon—195, 196, 902 and 951. Well 902, the most easterly of the four, is a few scores of feet from the alignment just inside the eastern border of the Badger Canyon watershed (Figure 1b). This well was initially impacted by tunnel construction in mid-January 2007. Distances from well 902 to the surface water sites in Badger Canyon watershed range from 0.27 miles to site 26, to almost 0.9 mile to site 152. Well 951 was also impacted initially in mid-January 2007. This well is less than one-fifth mile from the 27/213/214 site cluster, and more distant sequentially to sites 28, 26, 21, and 152. Wells 195 and 196, both initially impacted in mid-March 2007, are in—or close to—the middle portion of Badger Canyon. These two wells are 0.13 mile or less from sites 28 and 26, and progressively further distant from sites 21 and 152.

Site 214—Visual comparison of flows at site 214 and Ben Canyon reference site 11 do not suggest that a tunnel-related impact occurred at site 214 (Figures 109 and 110). Through 2003 flows at 214 typically were slightly less than those at 11 (Figure 110). In 2004 and 2005 flows at the two locations were approximately equal. Thereafter flows generally were slightly greater at 214 than at 11. The mid-January 2007 impact date for well 951, the closest well to the site 213/214/27 cluster, would suggest a mid-January 2007change in the flow relationship if an impact at 214 occurred. There is no change in the flow relationship between the two surface sites in early 2007. Furthermore, the baseflow comparison between sites 214 and 11 from 2007 through 2010 is well within the range of the pre-2007 baseflow relationship (Figure 111). Consequently there’s no evidence for a tunnel-related impact at site 214.

Site 213—Flows at site 213, in comparison to those at reference site 11, are similar to those at site 214. At 213 flows are equal to or slightly lower than those at 11 throughout the entirety of the period of record (Figures 112 and 113). And the baseflow comparison between sites 213 and 11 show no evidence of a tunnel-related impact to site 213; the 2007 through 2010 baseflow relationship between the two sites is within the range of the pre-2007 relationship (Figure 114). Consequently there’s no evidence for a tunnel-related impact at site 213.

Site 27—The impact assessment for site 27 is similar to those for sites 213 and 214. Flow at 27 ranges from slightly less to slightly more than at 11 (Figures 115 and 116) without a change at or near the anticipated early 2007 potential impact date. The baseflow comparison between sites 27 and 11 from 2007 through 2010 is also well within the range of the pre-2007 baseflow relationship (Figure 117). Consequently there’s no evidence for a tunnel-related impact at site 27.

Sites 28, 26, 21 and 152—These sites were not assessed for potential tunnel-related impacts for several reasons. Site 28 is at an ephemerally flowing location with insufficient (only nine) non-zero flow measurements to quantify either a baseline, pre-potential impact flow regime, or a potential post-impact regime. The last flow measurement at site 26, in late January 2003, is years before the earliest possible potential impact date. Similarly, the last reported measurement at site 21, in late August 2001, is years before the earliest possible potential impact date. Last, site 152 was monitored too infrequently (every six months prior to mid-2006) to allow adequate quantification of a potential pre-impact flow regime. Also the 6-month measurements, made typically each November and May, aren’t relevant to summer baseflow conditions.

Ben Canyon

Similar to many of the other ―smaller‖ watersheds in the project area, the south-southwest facing Ben Canyon watershed is bisected by the tunnel alignment (Figure 1b). Elevations in this catchment range from 3830 ft down to 1720 ft near percolation basins located in the alluvial plain west of the mountain front. At 284 acres area, the Ben Canyon watershed is close in size to Coldfoot Canyon watershed. A single main channel runs down the middle of the watershed, with no significant tributaries. Springsnails were found at two surface water sites in Ben Canyon (10 and 11), as was an adult Monterey Ensatina during surveys in 2002 and 2003.

The four surface water sites in the Ben Canyon watershed are all located within less than one-fifth mile of each other, and none of them are more than about one-tenth mile from the tunnel alignment (Figure 1b). Flow at site 10 (2485 ft elevation), the northernmost monitoring location, has been perennial throughout the monitoring period except during the summers of 2002 and 2003. Spikes in flow during winter suggest that flow at this site is affected by runoff. Summer flows are typically in the 1-6 gpm range. Site 157 (2440 ft elevation), a few hundred feet down-channel from site 157, flows at a lower rate than at site 10, with zero flow recorded most of the time between July 2000 and January 2005. Otherwise at site 157 flow has typically been less than 5 gpm. The flow regime at site 11 (2490 ft elevation), located on a minor tributary, is similar to that of site 10, with flow perennial except for the summers of 2002 through 2004. Springsnails have also been found at this location. TSite 9, the lowest site in the Ben Canyon watershed (2250 ft elevation), experienced reduced flow during the severe dry period from 2000 through 2004.

Wells 950 and 901 are both located in the central portion of the Ben Canyon watershed (Figure 1b). None of the surface water monitoring sites are greater than one-eighth mile for one of these two wells. Well 901 is located within 100 feet of the tunnel alignment, with well 950about one-sixth mile north of the alignment. Contrary to many other wells proximate to the alignment, water levels in neither of these two wells were impacted by tunnel construction. The lack of impact, particularly to well 901 almost co- located with the alignment, suggests that there may be no impacts to the surface water monitoring sites in Ben Canyon. If the surface sites were impacted, the time period for initial impact would likely be some time between the impacts to the closest impacted wells, 900 and 196. Well 196 was initially impacted in mid-March 2007, with well 900 initially impacted in mid-February 2008. Because the Ben Canyon surface sites are appreciably closer to well 900 than to well 196, the potential date of initial impact to the surface sites is estimated to be November 2007.

Site 10—Visual comparison of flows at site 10 with those at reference site 11 do not identify a significant change in the relationship between the two sites through the entirety of the monitoring record (Figures 118 and 119). Flows at site 10 are typically less than those at site 11 except for the late 2005 through late 2009 period when flow at 10 at times is slightly higher than at 11 (Figure 119). Presuming that a potential impact to site 10 could initially occur in November 2007, comparison of summer baseflows between sites 10 and 11 does not identify a changed relationship in 2007 or any later year versus flows in 2006 or earlier (Figure 120). The combination of a lack of documented well impact in the Ben Canyon watershed, plus no evidence of changed flow relationships between site 10 and reference site 11 means there’s no evidence for a construction-related impact at site 10.

Site 157—From either the visual comparison of flows for the duration of record for site 157 and reference site 11, or the summer baseflow comparison, there’s no evidence of a construction-related impact at site 157. Flows at 157 are always less than those at 11 (except when both are zero during the summers of 2002 through 2004) (Figures 121 and 122), and the summer baseflow relationship doesn’t change between the two sites in 2007, or thereafter, compared to the pre-2007 baseline relationship (Figure 123).

Site 11—Site 11 is used as a reference site in the assessment of many of the monitoring sites on the eastern tunnel alignment. To test the efficacy of using site 11 as a reference, flow comparisons were made between site 11 and reference site 185 in Sand Canyon. As with the other surface water monitoring sites in Ben Canyon, site 11 is located very close to well 901, and slightly more distant from well 950. Neither of these wells was impacted by tunnel construction, suggesting that the close-by surface monitoring sites also would be non-impacted. A good estimate of potential impact date for the Ben Canyon surface site cluster is February 2008, when the TBM is at its closest location to well 901. There’s no significant change in the flow relationship between sites 11 and 185 in late 2007 compared to earlier or later years (Figures 124 and 125). Flows were slightly higher at 185 than 11 from late 2006 through early 2007 (Figure 125), but this variation is not interpreted to be significant. Baseflow comparison between the two sites shows no evidence of a changed relationship; the relationship for each year from 2008 through 2011 is well within the baseline, pre-2008 loci of flows (Figure 126). Site 11 is therefore presumed to be un-impacted and can be used as a reference site.

Site 9—Similar to site 157, there’s no evidence for a construction-related impact at site 9. Flows at 9 were often equal to those at 11 through mid-2005, and again after mid-2009. Between these two periods flow at site 9 was actually slightly greater than at the reference site (Figures 127 and 128). Also the summer baseflow relationship doesn’t change between the two sites in 2007, or thereafter, compared to the pre-2007 baseline relationship (Figure 129).

Devil Canyon

Devil Canyon watershed is similar to City Creek watershed in that Devil drains a large area (over 3500 acres) ranging up to 5300 ft altitude on the San Bernardino Mountains plateau. As such Devil Canyon watershed differs from the smaller, lower elevation watersheds in receiving snow inputs during winter and a basically different hydrological regime. The tunnel curves to the southwest and ends at Devil Canyon portal at the southeastern corner of the Devil Canyon watershed (Figure 1b).

Four surface water monitoring locations in Devil Canyon watershed include three perennially flowing sites on private land on the main channel (193, 190, and 620), and a ―horizontal well‖ (site 110 at 2415 ft altitude) on a tributary about one-quarter mile east of the main channel. Two additional sites (8 and 153), located in small face watersheds between Devil and Ben Canyon, are included in this section (Figure 1b). Site 193, at 2310 ft altitude the highest of the three main channel sites, is located immediately below the junction of three forks of Devil Canyon Creek. Site 190, in the alluvium of the main channel at 2100 ft, and site 620, located on the gently sloping alluvial apron below the mountain front, at 1830 ft altitude, along with site 193, can have high winter flows—above 10,000 gpm—but with summer baseflows reaching zero at site 190 during drier years. Sites 8 (2055 ft altitude) and 153 (2000 ft altitude) are both located in ephemeral draws a few hundred feet from the base of the mountain front. Site 8 is a few hundred feet east of the tunnel alignment, and site 153 is about one-fifth mile east of the tunnel.

One well, 900, is relevant to the impact analysis of the surface water monitoring sites in the Devil Canyon area. 900 is located adjacent to the tunnel alignment a few hundred yards west of the western boundary of Ben Canyon (Figure 1b). This well was first impacted by tunnel construction in mid-February, 2008, and bottomed out two months later.

Site 193—Flow measurements at this site were made twice each year, typically in May and November. Consequently, there’s insufficient data to assess the potential for a tunnel-related impact at this site.

Site 110—Flow at site 110 was perennial through mid-June 2001, dry each summer through 2004, and thereafter perennial through the end of record in June 2008. The most likely time for a potential impact to begin at site 110 is early 2008, when well 900 was initially impacted. Because the data record at site 110 ends in June 2008, the surface water record is insufficient to determine the impact status of this site.

Site 190—The location of site 190, approximately 0.4 mile northwest of well 900, suggests that any tunnel-related impact at site 190 would begin no earlier than the impact at well 900, early 2008. Ocular comparison of flows at site 190 and reference site 11 through the entirety of the data record do not identify a shifted relationship before versus after 2008; flows at site 11 are about the same proportion of flows at site 190 throughout the entirety of the record (Figure 130) suggesting no impact at site 190. More detailed summer baseflow comparison between 190 and 11 reinforce the finding of no construction- related inpact at 190; although the baseline spread of blue diamonds is relatively wide in Figure 131, the data from 2008 through 2010 are well with the pre-2008/baseline relationship.

Sites 8 and 153—Flow at both of these sites has been minimal, with no flow recorded at site 8 from June 2005 through the last measurement in April 2011. At site 153 zero flow is reported during each summer of record except for 1998. Because there’s been essentially no reported baseflow at either of these sites, the baseflow comparison assessment technique is not applicable and the visual comparison

approach is minimally applicable at best. Consequently, it’s not possible to assess the impact status of these two sites.

Site 620—Site 620’s location southwest of the Devil Canyon portal implies that any construction- related impact at 620 would occur no earlier than the end of mining for the western tunnel, late July 2008. Unfortunately, data from only one summer, 2009, are available after July 2008. Consequently there are relatively few data points for basing an impact assessment. Also the baseline baseflow relationship is not ―tight‖, there’s a lot of spread in the pre-2009 points in Figure 133. Nevertheless, the three 2009 baseflow points are well within the range of the baseline baseflow relationship, and therefore there’s no reason to identify a tunnel-related impact at site 620.

IMPACTS TO TRIBAL LANDS

The USDA Forest Service has a legally-mandated tribal trust responsibility ―… to protect tribal lands, assets, resources, and treaty rights … . The trust responsibility is a legally enforceable obligation, a duty, on the part of the U.S. Government to protect the rights of Federally Recognized Indian Tribes‖ (p. 51, Section 2: Treaty Rights and Forest Service Responsibilities. www.fs.fed.us/people/tribal/trib-2.pdf accessed March 14, 2012). In this context staff of the San Bernardino National Forest are concerned about potential impacts to Tribal water and biological resources from potential Arrowhead tunnel construction.

The location of the San Manuel Tribe reservation at the foot of Sand Canyon means that potentially impacts upstream of the reservation on National Forest System lands in Sand Canyon, such as decreased water flows due to tunnel construction, could affect Tribal resources. Both the Tribe and MWD installed automated water gaging systems in lower Sand Canyon. These gages provided critical water flow data that would not otherwise have been available. In particular the gage data identified a wide daily/diel range of flow that the weekly MWD manual flow measurements in Sand Canyon did not identify. In summer, when evapotranspiration in Sand Canyon is high, flow magnitudes can vary over 30 gallons per minute (gpm) over a 24-hr period. Logistical considerations dictated that the manual, weekly flow measurements be made in late morning, coincidentally close to the time of day when flow is maximized. Consequently a 30 gpm manually-measured flow at site 117 could ―translate‖ to 0 gpm as the minimum daily flow. This is important because many aquatic biota need water to survive. When pools dry up (e.g., zero flow) biota could die. Tunnel construction effects did reduce flows at site 117 to within a range that temporarily and periodically could reduce daily minimum flows to zero. Automated measurements at the tribal gage confirmed this possibility, with consequent impacts to locations on the reservation. During several time periods irrigation was initiated specifically to keep pools and other channel stretches on the reservation from drying out. The tribe also initiated pool depth monitoring to help understand the flow dynamics on the reservation.

ENVIRONMENTAL EFFECTS TO SURFACE WATERS FROM ONGOING OPERATIONS AND MAINTENANCE OF THE ARROWHEAD TUNNELS PROJECT— As described in the 2012 Special Use Permit

The preceding text in this report addresses environmental effects associated with the construction of the Arrowhead Tunnels and lingering post-construction effects through summer 2012. This section deals with any anticipated environmental effects from operation and maintenance of the existing Arrowhead tunnels (e.g., the ―Proposed Action‖).

There are no anticipated environmental effects to surface waters in the project area anticipated with routine operation and maintenance of the Arrowhead Tunnels. The tunnels are steel-lined and consequently no groundwater should leak into the tunnels or otherwise be diverted from natural linkages to surface water sites. Any residual surface water effects from tunnel construction will not be completely determined until at least the end of summer 2012 but it is anticipated that at most monitoring of recovery of stream and spring flows would be needed through 2013 at locations in Sand Canyon and City Creek tributary sites 56/181/58. Because no direct or indirect effects are anticipated there will be no cumulative effects stemming from tunnel operation and maintenance.

A low likelihood possibility is that a seismic event of significant magnitude could rupture the tunnel or otherwise cause a breach in the tunnel lining allowing influx of groundwater into the tunnel. A breach could cause changes in the ground-surface water linkages and result in reduced surface flows. Although possible, this eventuality is considered to be remote enough to not require assessment for possible surface water effects. Furthermore, such an assessment would be coarse unless extensive time and effort were undertaken to model ground-surface water linkages at locations potentially susceptible to massive seismic activity. Conclusions from such a modeling effort would likely have a large error envelope. Consequently, for these reasons, no effort is made here to assess possible surface water impacts from one or more seismic events.

Because of a continuing groundwater impact in the vicinity of City Creek tributary sites 56, 58, and 181, monitoring of surface water here will be undertaken for up to five years in order to ensure the availability of surface water for wildlife. This monitoring is not associated with anticipated effects from the Proposed Action, rather it is due to a remnant groundwater drawdown from tunnel construction.

LESSONS LEARNED

Early on objectives for ecosystem health should be identified and understood by all parties. For instance, is the objective to retain ―average‖ or ―normal‖ functional levels, ―minimal‖ functionality, or something else? Clarity on the objective(s) allows the project proponent early on to understand his/her responsibilities, and allocate resources, including staff, accordingly.

Survey/inventory the project area for all surface water expressions and sensitive biotic species. Establish surface water monitoring sites at multiple locations within watersheds, to monitor representative tributaries and the main channel. Establish surface water monitoring sites at locations of known sensitive biota.

Assure adequate baseline data are available for surface water flow and groundwater levels representative of the complete project area. Ideally a 10-year pre-construction baseline is advisable. Failing an adequately long baseline, FS personnel should be extremely conservative and demand that the project proponent follow FS directives in the absence of an adequate baseline.

Identify early on potential types and characteristics of impacts to geologic, aquatic and riparian systems, such as -- o Lower groundwater table, potentially hundreds of feet? o De-water surface springs and streams, with consequent potential effects on biota—e.g., habitat loss and/or degradation, and actual loss of individuals of known species and potentially of un- described resident/transitory species?

o Realize that smaller, less obvious impacts may be difficult to quantify and therefore difficult to mitigate. o Duration of impact can vary, with shorter duration being harder to quantify and therefore difficult to mitigate. Groundwater drawdown could persist for years after tunnel is water-tight. o Potential for lag impacts/effects must be recognized and planned for (e.g., groundwater lowering may not ―translate‖ to surface water de-watering for weeks/months). o Realize that confounding factors like wildfire and climate change can complicate the identification, magnitude and timing of impacts.

Think through mitigation options way, way before a potential impact could occur.

Demand to have any mitigation procedures and equipment in place and tested before they are needed. Unexpected delays can occur, especially when mitigation is ―one of a kind‖ that the project proponent hasn’t done before. Unexpected glitches can occur that delay implementation. Consequently the need to have equipment in place, tested, and otherwise operational before needed is essential. Similarly, personnel need to be trained and knowledgeable of mitigation procedures before they are implemented. Aquatic biota can die if water in pools (for instance) is not available. This can happen over night—a very short time frame. Potential impacts, like pool drying, won’t wait for mitigation equipment to be operational and for crews to be functional.

Anticipate that there will be many critical details associated with mitigation and anticipate that all mitigation actions will take longer than expected.

Recognize the possibility of diel/diurnal water flow variation that could be substantial, and that could affect biota. Establish automated (near real-time) flow measurement sites as part of the baseline flow determination. Use automated surface flow measurements for determination of summer daily minimum flows that would not otherwise be quantified by ―spot‖, manual measurements. The manual measurements would probably be done at the time of day when flows are NOT at or near daily minimums. Site the automated gage(s) at a location(s) anticipated to be sensitive for species of interest, extent of riparian habitat, etc.

Manual/spot flow measurements less frequent than monthly are of very little use. Weekly spot measurements are needed when flows are naturally low (e.g., summer in Mediterranean climates) for timely response to impacts and to effectively mitigate.

As early in the project as possible, and prior to any tunnel construction, create a surface water flow model aimed at predicting anticipated natural flow magnitudes throughout the duration of the project. Use baseline data—potentially including precipitation, groundwater, and reference surface water monitoring information--as the input dataset.

CONCLUSIONS

Objectives of this assessment are to identify (1) surface water assess impacts (if any) from construction of the Arrowhead tunnels at the base of the San Bernardino Mountains between Devil Canyon Creek and City Creek, and (2) identify the level of recovery in surface waters from any impacts. The impacts could range from minor reductions in flows at streams and springs to complete de-watering of springs and/or stream reaches.

Construction of the Arrowhead tunnels inadvertently caused leakage of water from groundwater aquifers in the project area into the tunnels. This ―leaked‖ water would have potentially linked to the ground surface and became part of the surface water flow. Approximately 75 locations in the project area were monitored for surface water flow, typically beginning in 1994. Consequently a 17-year flow record was available, with often at least ten years of monitoring data available prior to any potential impact.

Two related procedures are used to assess potential construction-related impacts to surface waters. Both compare monitored flows at a reference (control) site with monitored flows at sites potentially impacted. The first procedure is an ocular comparison of flows through the entirety of data record. A visual change in the relationship between flows at the two sites at or about the time of documented construction impact at a proximate well(s) indicates a good likelihood for a construction-induced impact at the surface site. The second approach compares flows only for summer ―baseflow‖ periods when groundwater is the only source for surface water. Initially the relationship is generated for the pre-impact period (as determined by the date of impact at one or more nearby well(s)). A change from the pre-impact/baseline relationship indicates a high likelihood of a construction-induced impact to surface waters.

Flows at eighteen monitoring locations (including sixteen on NFS lands) were reduced at one time or another due to tunnel construction. In addition, flows on the San Manuel reservation at the base of Sand Canyon were also reduced. Mitigation, as irrigation at selected locations above many of the impacted sites, was routinely applied for several years. The irrigation effectively compensated for much of the construction-caused reductions in natural flows. As of summer 2012, flows at sites 48, 53, and 636, all in upper Sand Canyon, sites 56, 58, and 181 in a tributary to City Creek, and sites 45 and 154 in Borea Canyon continued to be barely within or below the anticipated natural flow ranges.

References

Ainsworth J, Doss TA 1995. Natural history of fire and flood cycles. California Coastal Commission. Accessed on line 5/18/11 at http://www.coastal.ca.gov/fire/ucsbfire.html.

Berg N 2008. Arrowhead Tunnel Project Irrigation Pipe Testing. Memo to the file available at San Bernardino National Forest Supervisor’s Office, San Bernardino, CA.

Cannon SH, DeGraff J 2009. The increasing wildfire and post-fire debris-flow threat in western USA, and implicatioins for consequences of climatic change. P. 117-190 in Sassa K and Canuti P (eds.) Landslides—Disaster Risk Reduction. Springer-Verlag, Berlin.

Gordon ND, McMahon TA, Finlayson BL 1992. Stream Hydrology—An introduction for ecologists. John Wiley & Sons. New York.

Lave J, Burbank D 2004. Denudation processes and rates in the Transverse Ranges, southern California: erosional response of a transitional landscape to external and anthropogenic forcing. J Geophy. Research 109:F01006.

Lisle TE, Cummins K, Madej MA 2007. An examination of references for ecosystems in a watershed context: results of a scientific pulse in Redwood National and State Parks, California. In ―Advancing the Fundamental Sciences: Proceedings of the Forest Service National Earth Sciences Conference, San Diego, CA, 18-22 October 2004‖. pp. 118-129 (US Department of Agriculture, Forest Service Northwest Research Station: Portland, OR).

Metropolitan Water District of Southern California (MWD) 2003. Arrowhead Tunnels Project Water Resource Monitoring and Mitigation Plan. On file at San Bernardino National Forest Supervisor’s Office, San Bernardino, CA.

Psomas 2004. Post Old Fire Evaluation of the Riparian Corridors Arrowhead Tunnel Alignments San Bernardino Mountains. Psomas 3187 Red Hill Ave., Suite 250, Costa Mesa, CA. On file at San Bernardino National Forest Supervisor’s Office, San Bernardino, CA.

Rantz SE et al. 1982. USGS Water Supply Paper 2175. Measurement and Computation of Streamflow. Volume 1. Measurement of stage and discharge.

Searcy JK 1959. Flow-duration curves; Manual of Hydrology: Part 2. Low-flow techniques. US Geology Survey Water-Supply Paper 1542-A.

US EPA 2007. An approach for using load duration curves in the development of TMDLs. EPA 841-B- 07-006. August 2007.

US Geological Service 2004. Ecoregions, reference conditions, and index calibration. Appendix F. Water. Available at http://acwi.gov/appendixes/AppendF.html.